@medicinejournal European Journal of Pediatric Surgery January 2020

Original Article
Potential Benefits of Laparoscopic Repair of

Duodenal Atresia: Insights from a Retrospective
Comparative Study
Martin Sidler1,2 Florin Djendov1 Joe I. Curry1 Simon Blackburn1 Stefano Giuliani1 Simon Eaton2
Dhanya Mullassery Kate M. Cross1 Paolo De Coppi1,2
1
1 Department of Specialist Neonatal and Paediatric Surgery, Great Address for correspondence Paolo De Coppi, MD, PhD, Stem Cells
Ormond Street Hospital for Children, London, United Kingdom and Regenerative Medicine Section, UCL Great Ormond Street
2 Stem Cells and Regenerative Medicine Section, UCL Great Ormond Institute of Child Health, 30 Guilford Street, London WC1N 1EH,
Street Institute of Child Health, London, United Kingdom United Kingdom (e-mail: [email protected]).
Eur J Pediatr Surg
Abstract Introduction Congenital duodenal obstruction (CDO) repair can be performed open
or laparoscopically. We aimed to determine the potential benefit of laparoscopic repair
regarding tolerance of enteral feeding, postoperative pain, hospital stay, and compli-
cation rate.
Materials and Methods In a single-center retrospective cohort study, we compared
neonates with isolated CDO operated open versus laparoscopically from 2010 to 2019.
No transanastomotic tubes were used, and anastomoses were created in a side-to-side
fashion in all cases. An early feeding policy is applied for all cases operated at our
institution. Statistical comparison was performed using the Mann–Whitney’s test or
Fisher’s exact test where appropriate.
Results Forty-one patients analyzed were similar regarding body weight, gestational
age, and proportion of patients with trisomy 21. Median follow-up was 21 months. Four
(20%) out of 20 laparoscopic procedures started laparoscopically were converted to
open. Comparing the 21 open with the 16 laparoscopically completed patients, median
anesthetic duration was shorter by 18% in the open versus laparoscopic completed
group (218 vs. 179 minutes, respectively; p ¼ 0.025). Median postoperative time to full
enteral feeds was shorter by 4 days in the first group (7 vs. 11 days, respectively;
p ¼ 0.028). In accordance, the median duration of parenteral nutrition (PN) was less
than half in the laparoscopic completed compared with the open group (5 vs. 11.5
Keywords days, respectively; p ¼ 0.031). Postoperative opioids were required for only half the
► neonatal duration in the laparoscopically completed group compared with open (2 vs. 4 days,
► surgery respectively; p ¼ 0.026). Outcomes such as length of stay, the occurrence of strictures
► duodenal atresia or adhesions requiring reintervention, or line sepsis were similar in both groups.
► laparoscopy Conclusion Patients undergoing laparoscopic CDO repair at our institution benefited
► minimally invasive from shorter time to full enteral feeds, and reduced the need for PN as well as
surgery postoperative pain medication.
received © Georg Thieme Verlag KG DOI https://doi.org/

May 15, 2019 Stuttgart · New York 10.1055/s-0039-1698766.
accepted after revision ISSN 0939-7248.
September 2, 2019
Potential Benefits of Laparoscopic Repair of Duodenal Atresia Sidler et al.
Introduction tion to transfer to neonatal intensive care was used. Given that
CDO repair is aimed at enabling the infant to feed normally,
The first series of repaired intrinsic congenital duodenal postoperative time to full enteral feeds was our primary
obstruction (CDO) was presented in 1931.1 Reports of success- outcome. Secondary outcomes were the duration of postoper-
ful minimally invasive surgery (MIS) to repair duodenal atresia ative ventilation, use of PN, need for opioids (i.e., morphine
(DA) were published 70 years later.2 Only a few reports have sulfate), and postoperative length of stay. We also recorded any
directly compared MIS with the traditional open approach; complications, such as anastomotic stricture or bowel adhe-
controversy still exists about the value of minimally invasive sions, requiring reoperation and instances of line infection.
CDO repair.3–7 On one hand, the latter may be beneficial in
terms of earlier feeding, less postoperative pain, earlier dis- Statistical Analysis
charge, and better cosmetic outcome. On the other hand, MIS Comparison of the preoperative patient characteristics
for CDO can be technically challenging and leads to longer included all patients in the Open and Lap groups; however,
procedures than open repair. Furthermore, it is unclear wheth- in the analysis of the peri- and postoperative data, only
er the rates of anastomotic leakage or stricture are higher after completely laparoscopic operated patients were included in
MIS compared with open surgery.8 Conversely, bowel adhe- the Lap group. To assess whether removing the laparoscopic-
sions, and associated adhesional obstruction, might plausibly converted-to-open patients significantly affected the results,
be more common after open repair.6,9 we also did a sensitivity analysis by grouping the patients as
We retrospectively review a single-center experience and per their intention to treat. Statistical analysis was performed
test the hypothesis that patients undergoing laparoscopic using IBM SPSS Statistics 25 (SPSS Inc., Chicago, Illinois, United
repair of CDO tolerate gastric feeds earlier, have a corre- States). Due to the sample size and since most variables were
spondingly decreased need for parenteral nutrition (PN), not normally distributed (as assessed using the Shapiro–
need less postoperative pain medication, and have a shorter Wilk’s test), we compared continuous factors and outcome
hospital stay. variables using the Mann–Whitney’s test and dichotomous
factors using the Fisher’s exact test. Numerical results are
graphically presented using boxplots, whereby the box
Materials and Methods
encompasses the interquartile range from the first quartile
This was a single-center, retrospective review of patients to the third quartile, and the bold transverse bar representing
undergoing open or MIS repair of CDO. The study was the median. The whiskers mark maximum and minimum
approved by the institution’s audit and governance board values. Outliers represented by a circle () or an asterisk () are
(registration number 2510). more than a 1.5-fold or a 3-fold interquartile range away from
quartile 1 or 3, respectively. Probability values of 0.05 were
Patients considered significant, whereas values >0.05 were considered
Patients younger than 2 weeks undergoing open or laparoscopic nonsignificant (NS).
repair of CDO between March 2010 and March 2019 were
considered for the study. Exclusion criteria were concomitant
Results
anorectal malformation, esophageal atresia, or major cardiac
anomalies with only palliative treatment options. Patients During the study period, 50 patients of less than 2 weeks
either underwent laparoscopic (Lap group) or open (Open postnatal age underwent repair of CDO. Patients with con-
group) repair of CDO decided based on anticipated tolerance comitant esophageal atresia (five patients) or anorectal
of a pneumoperitoneum and the operating surgeon’s prefer- malformation requiring a colostomy (three patients), as
ence. The group of patients having their repair completed well as one patient with tricuspid atresia and large ventricle
laparoscopically is referred to as Lap completed group. For septum defect who underwent DA repair for palliation only,
both the open and the laparoscopic approaches, the incisions were excluded from the study.
in the proximal and distal duodenum were oriented such as they
permitted to do a side-to-side anastomosis, completed with Patient Characteristics
single interrupted stitches of fine monofilament suture materi- Twenty-one patients in the Open group and 20 patients in
al. No transanastomotic tubes (TATs) were placed. Laparoscopic the Lap group were similar in terms of their weight, age,
setup comprised an umbilical camera port, two working ports, prematurity, trisomy 21, presence of intestinal malrotation,
and a Nathanson retractor insertedjust under the xiphisternum. cardiac anomalies, and gender distribution (►Table 1).
Perioperative Factors and Outcome Perioperative Data and Outcome

Data collected were weight, postnatal age, prematurity, gender, Intestinal malrotation required a Ladd’s procedure in 3 of 21
chromosomal anomalies, presence of significant cardiac anom- patients in the Open group and 2 of 16 patients in the Lap
alies (such as large ventricular septal defect, tetralogy of Fallot, completed group (NS). Median duration of anesthesia was
or double outlet right ventricle), additional gastrointestinal shorter by 18% in the Open group compared with the Lap
conditions (such as malrotation), or further small bowel completed group (179 vs. 218 minutes, range 133–274 vs. 155–
atresias. Since the duration of the operation was only available 389 minutes, respectively; p ¼ 0.025; ►Fig. 1); we did not
for a minority of patients, duration of anesthesia from induc- observe a significant change in anesthetic duration over
European Journal of Pediatric Surgery

Table 1 Preoperative patient characteristics Table 2 Perioperative data and outcome
Primary open Laparoscopic p-Value Primary open Laparoscopic p-Value

repair, n ¼ 21 repair, n ¼ 20 repair, n ¼ 21 completed
repair, n ¼ 16
Weight at 2.64 2.67 0.76
operation (kg)a (1.1–4.0) (1.1–3.6) Concomitant 3 (14.3%) 2 (12.5%) > 0.99
Ladd’s
Age (days)a 2 (0–8) 2 (0–14) 0.81
procedure
Prematurityb 8 (38%) 7 (35%) > 0.99
Postoperative 1 (0.5–5) 1 (0.5–2) 0.63
Trisomy 21 6 (29%) 5 (25%) > 0.99 invasive
ventilation
Cardiac 5 (24%) 6 (30%) 0.73
(days)a
anomaly
Line infection 3 (14.3%) 2 (12.5%) > 0.99
Male:Female 11:10 8:12 0.54
ratio Postoperative 16.5 (5–27) 13.5 (5–58) 0.94
length of stay
Intestinal 3 (14.3%) 2 (10%) > 0.99
(days)a
malrotation
Anastomotic 2 (9.5%) 2 (12.5) > 0.99
a
Data reported as median and range. stricture
b
Born before 37 weeks’ gestation. requiring
procedure
a
Data reported as median and range.
time. Six of 10 patients with anesthetic durations of 4 hours or
more had cardiac anomalies such as double outlet right ven-
tricle or tetralogy of Fallot. One patient also had a laparoscopic (7 vs. 11 days, range 4–24 vs. 4–18 days, respectively;
Ladd’s procedure, which added to the duration of surgery. The p ¼ 0.028; ►Fig. 2). In accordance, the median duration of
median postoperative duration of treatment with opioid med- PN was less than half in the Lap completed group compared
ication, on the other hand, was half the duration in the Lap with the Open group (5 vs. 11.5 days, range 0–40 versus 0–25
completed group compared with the Open group (2 vs. 4 days, days, respectively; p ¼ 0.031). The rate of line infections and
range 1–14 vs. 0–5 days, respectively; p ¼ 0.026; ►Fig. 1). The the rate of anastomotic strictures requiring a further duodenal
duration of postoperative ventilation was similar in both procedure were similar in the Open group compared with the
groups (median of 1 day in either group; NS) (►Table 2). Lap completed group (two in both groups and three [14.3%]
The median postoperative time to reaching full enteral patients vs. two patients [12.5%], respectively). One patient in
feeds was shorter by 4 days in the Lap completed group the Lap completed group had a localized adhesiolysis prior to a
Fig. 1 Laparoscopic repair was associated with decreased need for opioids yet also with longer procedure times. (A) Median duration of
postoperative need for opioid pain medication was half in patients after a laparoscopic repair compared with open (2 vs. 4 days, respectively;
#
p ¼ 0.026). (B) Median anesthetic duration was 39 minutes longer in patients operated laparoscopically compared with open (218 vs.
179 minutes, respectively; § p ¼ 0.025).

Fig. 2 Earlier full enteral feeding and decreased need for parenteral nutrition (PN) after laparoscopic repair. (A) Median time to full feeds (i.e.,
gastric or oral) was 36% shorter in the laparoscopic group compared with open (7 vs. 11 days, respectively; # p ¼ 0.028). (B) Median duration of PN
was less than half after laparoscopic repair compared with open (5 vs. 11.5 days, respectively; §p ¼ 0.031).
revision of the anastomosis, which did not sufficiently improve Median times to full enteral feeding and duration of PN were
passage through the duodenum to resolve symptoms. shorter by 4 and 5.5 days, respectively (7 vs. 11 days,
There was a tendency to a shorter stay in the Lap com- respectively; p ¼ 0.012 and 6 vs. 11.5 days, respectively;
pleted group, with a median postoperative length of stay of p ¼ 0.28 comparing the Lap group with the Open group.
13.5 versus 16.5 days in the Open group (NS). In the Lap Median duration of postoperative invasive ventilation and
completed group, four patients stayed longer than the medi- length of postoperative hospital stay was similar in the Lap
an time because discharge was delayed by the fact that they group compared with the Open group (0.5 vs. 1 day, respec-
were awaiting review by another specialty. One patient was tively; p ¼ 0.50 and 11 vs. 16.5 days, respectively; p ¼ 0.54).
awaiting transfer to a local hospital to complete transition
from nasogastric tube feedings to oral feeds, whereas the
Discussion
vast majority of our patients were discharged home on oral
feeds. Also, in the Lap completed group, one patient required Treatment of CDO can be performed using an open operation
reintervention due to anastomotic stricture, while another or via laparoscopy. Controversy still exists about the risks
patient was kept longer due to poor weight gain. The reasons and benefits of either approach; only a few studies have
for prolonged hospital stay were different in the Open group, directly compared the open versus the MIS technique.3–7
where 7 out of 10 patients staying longer than median In our analysis, the overall anesthetic time from induction
postoperative time to discharge were in fact kept because to transfer out of the operating room was longer in the
they had not reached full enteral feedings. Median follow-up laparoscopic patients. Significantly, longer surgical procedure
was 21 months. times for MIS repair were also found by three out of
Of the four patients converted from laparoscopy to an five comparative reports.3–7 Additionally, cardiac comorbid-
open operation, three patients had a duodenal web such that ities may have contributed to the prolonged anesthetic dura-
the anatomy could not be reliably defined; the other patient tion in some laparoscopic cases.
had a complete DA, weighed 1.13 kg, and did not tolerate the Tolerance of full enteral feedings is undoubtedly the main
pneumoperitoneum. goal of CDO repair. In our experience, time to full feeding
In an intention-to-treat analysis, we also compared the after laparoscopic surgery was clearly shorter than after
outcome of the Open group with the Lap group, which open surgery; our patients’ times to full enteral feeds were
included the four procedures that were converted to an similar or even markedly shorter relative to other reports
open operation. Median anesthesia time was 14% shorter comparing open with MIS CDO repair.3,5–7 Correspondingly,
in the Open group compared with the Lap group (179 vs. the duration of PN was also reduced after laparoscopic repair.
208 minutes, respectively; p ¼ 0.024). Median duration of While PN is known to increase the risk of cholestasis,
postoperative opioid requirements was shorter in the Lap thrombosis, and the rate of sepsis,10,11 we were unable to
group compared with the Open group, although this differ- demonstrate a benefit of MIS regarding the occurrence of
ence was not significant (2 vs. 4 days, respectively; p ¼ 0.18). line infections in our patient cohort. Interestingly, our

experience is similar to other reports when considering MIS, operations are currently available,6 many accept that lapa-
but it is superior to others when considering the open roscopy leads to fewer adhesions than laparotomy.9
approach in that our patients who had open surgery reached This present report as well as other studies comparing
full feeds sooner than in other published series.8 This could open versus laparoscopic repair of CDO needs to be inter-
be related to the early feeding policy, and we have been preted with caution due to their retrospective methodology.
following with small intermittent nasogastric feeds starting Ideally, the apparent benefits of laparoscopic CDO repair
on or around the second postoperative day. Our policy would be confirmed in a randomized controlled trial. How-
following this early commencement of enteral feeds is simi- ever, given the rarity of the condition, it is uncertain whether
lar to that described by others.12 This practice may well a randomized controlled trial is feasible. We also analyzed
contribute to our experience of early tolerance of gastric our data on an “intention-to-treat” basis, as in a trial, and the
feeds; the studies describing initial feeds around median results suggest that overall, it still seems beneficial to start
days 5.4 to 11 report longer times to reach full feeds after the operation laparoscopically, although 20% of infants
laparoscopic as well as open surgery.3–6 Because of our early would be expected to require conversion to an open
feeding policy, we have not adopted the use of a TAT. Placing a procedure.
TAT is another means by which enteral (i.e., postanasto-
motic) feeds can be advanced more quickly after surgery.13,14
Conclusion
While they might reduce the need for PN, however, TATs have
not been proven to reduce time to full preanastomotic feeds Patients with CDO operated laparoscopically appear to ben-
or to reduce the length of hospital stay.15 The reported efit from a reduced postoperative need for opioids, earlier
median times to full postanastomotic feeds are similar to full feeds, and a reduced need for PN. The duration of MIS
our findings for full oral and gastric feeds after MIS repair13; repair seems to be longer than open surgery.
placing a TAT tube has not been shown to reduce time to full
preanastomotic feeds.14 Furthermore, the potential benefits Funding
of a TAT need to be weighed carefully with its associated risks M.S. and P.D.C. are supported by National Institute for
such as displacement or perforation,13,16 which should be set Health Research (NIHR-RP-2014-04-046). All researches
against the potential to reduce the rate of line sepsis13 and at the Great Ormond Street Hospital NHS Foundation
PN-related cholestasis.11 Trust and UCL Great Ormond Street Institute of Child
Despite the clearly earlier feeding after laparoscopic Health are made possible by the NIHR Great Ormond
repair compared with open surgery in our cohort, the Street Hospital Biomedical Research Centre. The views
postoperative length of hospital stay was not significantly expressed are those of the author(s) and not necessarily
shorter in the first group. In our laparoscopic patients, the those of the NHS, the NIHR, or the Department of Health.
reasons prolonging their hospital stay were predominantly
logistic in nature and unrelated to their tolerance of gastric or Conflict of Interest
oral feeds. None declared.
There seems to be various reliable techniques to perform a
safe duodenal anastomosis. One group found Nitinol U-clips to
work well,3 while others found running sutures to markedly References
1 Ladd WE. Congenital obstruction of the duodenum in children.
reduce the risk of an anastomotic leak8; both report formation
N Engl J Med 1932;206:277–283
of a diamond-shaped anastomosis. However, we have been 2 Bax NM, Ure BM, van der Zee DC, van Tuijl I. Laparoscopic
using a side-to-side anastomosis for both, the open and the duodenoduodenostomy for duodenal atresia. Surg Endosc
laparoscopic techniques.10 The duodenal incisions placed 2001;15(02):217–219
transversely on the proximal segment and longitudinally on 3 Spilde TL, St Peter SD, Keckler SJ, Holcomb GW III, Snyder CL,
the distal one allow us to complete our anastomoses in a side- Ostlie DJ. Open vs laparoscopic repair of congenital duodenal
obstructions: a concurrent series. J Pediatr Surg 2008;43(06):
to-side fashion, with single interrupted stitches of fine mono-
1002–1005
filament sutures, either absorbable or nonabsorbable. This 4 Hill S, Koontz CS, Langness SM, Wulkan ML. Laparoscopic versus
technique has not been modified by the laparoscopic approach open repair of congenital duodenal obstruction in infants. J Lapa-
which makes the data presented here comparable. We specu- roendosc Adv Surg Tech A 2011;21(10):961–963
late that one patient in our Lap completed group may have had 5 Jensen AR, Short SS, Anselmo DM, et al. Laparoscopic versus open
treatment of congenital duodenal obstruction: multicenter short-
a small leak leading to localized adhesions and anastomotic
term outcomes analysis. J Laparoendosc Adv Surg Tech A 2013;23
stricture, although this patient did not have any other clinical (10):876–880
signs of this. 6 Cho MJ, Kim DY, Kim SC, Namgoong JM. Transition from laparot-
The risk of small bowel adhesions is another important omy to laparoscopic repair of congenital duodenal obstruction in
factor when considering the type of approach. Postoperative neonates: our early experience. Front Pediatr 2017;5:203
small bowel adhesions can cause obstruction, chronic pain, 7 Son TN, Kien HH. Laparoscopic versus open surgery in manage-
ment of congenital duodenal obstruction in neonates: a single-
and even female infertility. Reported rates of small bowel
center experience with 112 cases. J Pediatr Surg 2017;52(12):
adhesions leading to obstruction after laparotomy for DA are 1949–1951
around 11% and after laparotomy for intestinal atresia as high 8 van der Zee DC. Laparoscopic repair of duodenal atresia: revisited.
as 28%.17 While only limited data after neonatal laparoscopic World J Surg 2011;35(08):1781–1784

9 Gutt CN, Oniu T, Schemmer P, Mehrabi A, Büchler MW. Fewer venous access and parenteral nutrition in infants with congenital
adhesions induced by laparoscopic surgery? Surg Endosc 2004;18 duodenal obstruction. Pediatr Surg Int 2011;27(08):851–855
(06):898–906 14 Harwood R, Horwood F, Tafilaj V, Craigie RJ. Transanastomotic tubes
10 Bishay M, Lakshminarayanan B, Arnaud A, et al. The role of reduce the cost of nutritional support in neonates with congenital
parenteral nutrition following surgery for duodenal atresia or duodenal obstruction. Pediatr Surg Int 2019;35(04):457–461
stenosis. Pediatr Surg Int 2013;29(02):191–195 15 Ruangtrakool R, Mungnirandr A, Laohapensang M, Sathornkich C.
11 Jiang W, Lv X, Xu X, Geng Q, Zhang J, Tang W. Early enteral Surgical treatment for congenital duodenal obstruction. J Med
nutrition for upper digestive tract malformation in neonate. Asia Assoc Thai 2001;84(06):842–849
Pac J Clin Nutr 2015;24(01):38–43 16 Sun SC, Samuels S, Lee J, Marquis JR. Duodenal perforation: a rare
12 Ekingen G, Ceran C, Guvenc BH, Tuzlaci A, Kahraman H. Early complication of neonatal nasojejunal tube feeding. Pediatrics
enteral feeding in newborn surgical patients. Nutrition 2005;21 1975;55(03):371–375
(02):142–146 17 Fredriksson F, Christofferson RH, Lilja HE. Adhesive small bowel
13 Hall NJ, Drewett M, Wheeler RA, Griffiths DM, Kitteringham LJ, obstruction after laparotomy during infancy. Br J Surg 2016;103
Burge DM. Trans-anastomotic tubes reduce the need for central (03):284–289

Original Article
“Tuft Cells”: A New Player in Hirschsprung’s

Disease
Anne Marie O’Donnell1 Hiroki Nakamura1,2 Prem Puri1
1 National Children’s Research Centre, Our Lady’s Children’s Hospital, Address for correspondence Anne Marie O’Donnell, PhD, National
Crumlin, Ireland Children’s Research Centre, Our Lady’s Children’s Hospital, Crumlin,
2 Department of Pediatric General and Urogenital Surgery, Juntendo Ireland (e-mail: [email protected]).
University, Tokyo, Japan
Eur J Pediatr Surg
Abstract Introduction “Tuft” cells, also known as brush or caveolated cells, are characteristi-
cally fusiform shaped, with a distinct apical “tuft” of microvilli extending into the
lumen. Double cortin-like kinase 1 (DCLK1) is a microtubule kinase and is a specific
marker of intestinal tuft cells. DCLK1-positive tuft cells have been shown to play a key
role in gastrointestinal chemosensation, inflammation, and neurotransmission. DCLK1
and Choline acetyltransferase (ChAT), the enzymes responsible for acetylcholine
production, are reported to be coexpressed within the gastrointestinal tract. We
designed this study to investigate the hypothesis that DCLK1 gene expression is altered
in Hirschsprung’s disease (HSCR).
Materials and Methods HSCR tissue specimens (n ¼ 6) were collected at the time of
pull-through surgery, while control samples were obtained at the time of colostomy
closure in patients with imperforate anus (n ¼ 6). Quantitative real-time polymerase
chain reaction (qRT-PCR) analysis was undertaken to quantify DCLK1 gene expression,
and immunolabeling of DCLK1-positive tuft cells was visualized using confocal
microscopy.
Results qRT-PCR analysis revealed significant downregulation of the DCLK1 gene in
both aganglionic and ganglionic HSCR specimens compared with controls (p < 0.05).
Keywords Confocal microscopy revealed DCLK1-positive tuft cell expression within the colonic
► tuft cells mucosa, with a reduction in expression in both aganglionic and ganglionic HSCR colon
► double cortin-like compared with controls.
kinase 1 Conclusion DCLK1 is significantly downregulated in HSCR colon, suggesting a role for
► Hirschsprung’s tuft cells in cholinergic neurotransmission of the distal colon. The marked decrease in
disease DCLK1 expression within ganglionic specimens highlights the physiologically abnormal
► aganglionosis nature of this segment in HSCR patients.
Introduction been shown to play a key role in gastrointestinal chemo-

sensation, inflammation, and intestinal repair after injury.1,2
“Tuft” cells, also known as brush or caveolated cells, are A recent paper by Leppänen et al documented the expres-
characteristically fusiform shaped, with a distinct apical sion of DCLK1-positive enterocytes in the human intestine
“tuft” of microvilli extending into the lumen. Double cortin- for the first time.3 Tuft cells were found along the entire
like kinase 1 (DCLK1) is a microtubule kinase and is a specific length of the human gastrointestinal (GI) tract. These cells
marker of intestinal tuft cells. DCLK1-positive tuft cells have were seen to be scattered between the crypts and villi, with

September 17, 2019
New Player in Hirschsprung’s Disease: Tuft Cells O’Donnell et al.
DCLK1 expressed on the basal pole of each cell. They did not 30 minutes to avoid nonspecific absorption, sections were
show morphological features typical to tuft cells, as reported incubated with primary antibodies; rabbit anti-DCLK1
in animal studies,1,2 but instead showed an apical brush (Abcam, United Kingdom), mouse anti-Chat (Santa Cruz),
border of even thickness.3 In the duodenum, they found mouse anti-HuC/D (Santa Cruz), all used at a dilution of
colocalization of synaptophysin and DCLK1, suggesting a 1:100 in PBS þ 0.05% TritonX-100, overnight at 4°C. Sections
neuroendocrine role for these cells in this region. were then washed in PBS þ 0.05% Tween and incubated with
Hirschsprung’s disease (HSCR) is a congenital disease of corresponding secondary antibodies; goat anti-rabbit Alexa
the newborn, characterized by the absence of ganglion cells Fluor 488, dilution 1:200 and goat antimouse Alexa Fluor 594,
in the distal colon. This condition has an incidence of 1,500 dilution 1:200, Abcam, United Kingdom, for 1 hour at room
live births with a male to female preponderance of 4:1. temperature. After washing, sections were counterstained
Treatment for HSCR involves the surgical resection of the with DAPI (4′,6-diamidino-2-phenylindole) antibody, dilution
affected aganglionic segment. Despite a properly performed 1:1,000 (Roche Diagnostics GmbH, Mannheim, Germany) for
pull-through operation, many patients (35–48%) unfortu- 10 minutes, washed, mounted, and coverslipped with Fluores-
nately suffer from persistent dysmotility symptoms, such as cent Mounting Medium (DAKO Ltd, Cambridgeshire, United
constipation, soiling, and an inflammatory condition of the Kingdom). All sections were independently evaluated by two
bowel known as enterocolitis.4–6 Many theories exist regard- investigators with a LSM 700 confocal microscope (Carl Zeiss
ing the reason why these patients continue to suffer post- MicroImaging GmbH, Jena, Germany).
surgery. In recent years, our group has published mounting
evidence that suggests that the “normal” ganglionic seg- Quantitative Real-Time Polymerase Chain Reaction
ment, which remains intact in HSCR patients postsurgery, is TRIzol reagent (Invitrogen) was used for the acid guanidinium-
actually remarkably abnormal.7–10 We designed this study to thiocyanate-phenol-chloroform extraction method to isolate
investigate the expression of DCKL1-labeled tuft cells in total RNA from HSCR and control tissues (n ¼ 6 for each group)
HSCR colon versus normal controls. according to the manufacturer’s protocol. Spectrophotometri-
cal quantification of total RNA was performed using a Nano-
Drop ND-1000 UV–Vis spectrophotometer (Thermo Scientific
Fisher, Wilmington, Delaware, United States). The RNA solu-
Tissue Samples tion was stored at 20°C until further use. cDNA synthesis and
This study was approved by the Ethics Medical Research quantitative polymerase chain reaction reverse transcription
Committee, Our Lady’s Children’s Hospital, Dublin, Ireland of total RNA was performed at 85°C for 3 minutes (denatur-
(Ref. GEN/292/12) and tissue samples were obtained with ation), at 44°C for 60 minutes (annealing), and at 92°C for
informed parental consent. HSCR specimens from six patients 10 minutes (reverse transcriptase inactivation) using a Tran-
who underwent pull-through surgery were studied. These scriptor High-Fidelity cDNA Synthesis Kit (Roche Diagnostics,
specimens were divided into aganglionic and ganglionic speci- West Sussex, United Kingdom) according to the manufac-
mens. We compared the most distal aganglionic segments turer’s instructions. The resulting cDNA was used for quanti-
with the most proximal ganglionic segments. HSCR patients tative real-time polymerase chain reaction (qRTPCR) using a
were aged 6 3 months old. No additional health issues LightCycler 480 SYBR Green-I Master (Roche Diagnostics,
existed in these patients. Colonic control samples included Mannheim, Germany) in a total reaction mix of 20 µl per
six specimens from patients who underwent colostomy clo- well. The following gene-specific primer pairs were used:
sure following surgical correction of imperforate anus. The Human DCLK1 (Eurofins) sense primer 5′ TGA ACG TCA AGA
normal control samples we use are taken 1.5 cm from the site CCA CCT CG and Human DCLK1 (Eurofins) antisense primer 5′
of anastomosis, during closure of the colostomy, which is ATC CTG ACA GCT TTC CGT GG. For normalization purposes,
generally considered normal colonic tissue. Control samples real-time RT-PCR was performed for glyceraldehyde 3-phos-
were taken from patients who were 11 4 months old. None of phate dehydrogenase (GAPDH). GAPDH sense primer 5′ACA
the imperforate anus patients had HSCR. Tissue specimens TCG CTG AGA CAC CT GG and GAPDH antisense primer 5′ GAC
were either snap-frozen in liquid nitrogen or stored at 80°C GGT GCC ATG GAA TTT GC were used. After 5 minutes of initial
for protein extraction or embedded in OCT Mounting Com- denaturation at 95°C, 55 cycles of amplification for each
pound (VWR International, Leuven, Belgium) for immunoflu- primer were performed. Each cycle included denaturation at
orescence and stored at 80°C until use. 95°C for 10 seconds, annealing at 60°C for 15 seconds, and
elongation at 72°C for 10 seconds. Relative mRNA levels of gene
Immunofluorescence Staining and Confocal expression were determined using a LightCycler 480 System
Microscopy (Roche Diagnostics) and the relative changes in gene expres-
Frozen blocks of HSCR colon and normal control samples were sion were normalized against the level of GAPDH gene expres-
sectioned transversely at a thickness of 10 µm, mounted on sion in each sample. Experiments were performed in duplicate
SuperFrost Plus slides (VWR International, Leuven, Belgium) for each sample and primer.
and fixed with 10% buffered formalin for 5 minutes. Sections
underwent cell membrane permeabilization with 1% TritonX- Statistical Analyses
100 for 20 minutes at room temperature. After blocking with A one-way analysis of variance (ANOVA) was conducted to
10% normal goat serum (Sigma Aldrich Ltd, Arklow, Ireland) for determine a statistically significant difference between

Fig. 1 Immunofluorescent staining of DCLK1-labeled tuft cells (green) (A), found to be co-localized with Choline acetyltransferase (ChAT; red)
(B), in the ganglionic HSCR colon. Nuclei were stained with DAPI (blue). Arrows show colocalisation. DCLK1, Double cortin-like kinase 1; HSCR,
Hirschsprung’s disease.
aganglionic, ganglionic and healthy controls (p < 0.05). Data decreased in both the aganglionic and ganglionic HSCR colon
are presented as mean standard error. Specimens were compared with normal controls (►Fig. 1). DCLK1-positive
classified into three groups: aganglionic (n ¼ 6), ganglionic cells were seen as elongated cells distributed in a scattered
(n ¼ 6), and healthy controls (n ¼ 6). pattern between the crypts and villi of the colon.
Quantitative Real-Time Polymerase Chain Reaction

Results
The relative mRNA expression level of the DCKL1 gene was
Immunofluorescence Staining significantly decreased in both aganglionic and ganglionic
Imunofluorescence in conjunction with confocal microscopy HSCR specimens compared with normal controls (p < 0.05;
revealed that the distribution of DCLK1-positive cells was ►Fig. 2)
Fig. 2 qRT-PCR revealed significantly decreased relative mRNA expression levels of DCLK1 in both the aganglionic and ganglionic HSCR
specimens (n ¼ 6) compared with normal control tissue (n ¼ 6). Results are presented as mean SEM (p < 0.05, p < 0.001). DCLK1, Double
cortin-like kinase 1; HSCR, Hirschsprung’s disease; qRT-PCR, quantitative real-time polymerase chain reaction; SEM, standard error of the mean.

Discussion tracts of mice express an incomplete neuronal-type cholin-

ergic phenotype, with presence of ChAT, but lack vesicular
Tuft cells were initially described in the murine stomach and acetylcholine transporter (VAChT) and choline transporter 1
duodenum. They have a unique fusiform shape and a distinct (CHT1) that are usually vital in fully functional cholinergic
apical “tuft” of microvilli extending into the lumen. Tuft cells neurons.1 It is thus speculated that ACh is released from
belong to an epithelial lineage distinct from goblet, enter- these cells in a different manner.
oendocrine, absorptive, and Paneth cells,2 and are identified
histologically by their strong expression of DCKL1. They have
Conclusion
been shown to trigger interleukin-25 associated immune
reactions leading to expulsion in pathogens and to be In this current study, we have shown that DCLK1 is signifi-
involved in epithelial protection and regeneration.11 The cantly downregulated in HSCR colon, suggesting a role for
colonic epithelium is renewed every 4 to 5 days in humans.2 tuft cells in cholinergic neurotransmission of the distal colon.
Physiological maintenance of the colonic epithelium is de- The marked decrease in DCLK1 expression within ganglionic
rived from colonic epithelial stem cells. Tuft cells have also specimens highlights the physiologically abnormal nature of
been shown to expand dramatically in response to chronic this segment in HSCR patients. We believe that these issues
inflammation, particularly in the early stages of carcinogen- are partly responsible for the bowel dysmotility experienced
esis that has led to the notion that they might also function as by many HSCR patients, following a properly performed pull-
cancer-initiating cells.12 through operation. It is no longer sufficient to use the
The crosstalk between tuft cells and neurons has been presence of ganglia as a marker of normal bowel in the
studied by numerous research groups. In one mouse study, ganglionic region. Tuft cell expression could potentially be
60% of tuft cells in the small intestine and 40% in the large used as a marker of normal bowel.
intestine were found to be in contact with nerve fibers.11
Calcitonin gene-related peptide-immunoreactive (CGRP)- Funding
nerve fibers constituted one third of the nerve fiber contacts This study was funded by Health Research Board (grant/
in the small intestine and two-thirds in the large intestine. award number: 4RA-BR-2015-1155).
CGRP is a multifunctional neurotransmitter involved in
maintaining the intestinal epithelial barrier by controlling Conflict of Interest
stem/progenitor cells and thus proliferation.11 This data None declared.
suggest a role for these cells in the modulation of GI activities
in response to luminal signaling.11 Cheng et al have recently
reported the coexpression of serotonin by tuft cells. They References
1 Schütz B, Jurastow I, Bader S, et al. Chemical coding and chemo-
found that in the murine small intestine, up to 80% of all tuft
sensory properties of cholinergic brush cells in the mouse gas-
cells also expressed serotonin, while in the large intestine trointestinal and biliary tract. Front Physiol 2015;6:87
such cells were rare.13 The highest number of DCLK1 cells 2 Yi J, Bergstrom K, Fu J, et al. Dclk1 in tuft cells promotes
coexpressing serotonin was found in the upper small intes- inflammation-driven epithelial restitution and mitigates chronic
tine with numbers gradually decreasing distally. The authors colitis. Cell Death Differ 2019;26(09):1656–1669
concluded that these constituted a novel subset of tuft cells 3 Leppänen J, Helminen O, Huhta H, et al. Doublecortin-like kinase
1-positive enterocyte - a new cell type in human intestine. APMIS
that may have a unique role in the GI tract.13
2016;124(11):958–965
Westphalen et al have shown that in vitro experiments 4 Menezes M, Corbally M, Puri P. Long-term results of bowel
using organoids from the gut revealed that under standard function after treatment for Hirschsprung’s disease: a 29-year
culture conditions, tuft cells do not survive in gut organo- review. Pediatr Surg Int 2006;22(12):987–990
ids.14 They found that tuft cells were sustained only when 5 Menezes M, Pini Prato A, Jasonni V, Puri P. Long-term clinical
outcome in patients with total colonic aganglionosis: a 31-year
cocultured with neurons, thus representing the first cell type
review. J Pediatr Surg 2008;43(09):1696–1699
that cannot be generated in organoid cultures in the absence
6 Laughlin DM, Friedmacher F, Puri P. Total colonic aganglionosis: a
of nonepithelial niche signals. Consistent with those find- systematic review and meta-analysis of long-term clinical out-
ings, the same authors found that tuft cells were largely come. Pediatr Surg Int 2012;28(08):773–779
absent in Ret/ mice, which display an aganglionic color- 7 O’Donnell AM, Coyle D, Puri P. Decreased expression of hyper-
ectum, as well as in extrinsically denervated intestine fol- polarisation-activated cyclic nucleotide-gated channel 3 in
Hirschsprung’s disease. World J Gastroenterol 2015;21(18):
lowing small intestinal transplantation.14 HSCR colon
5635–5640
displays aganglionosis, thus the results of our current study 8 O’Donnell AM, Coyle D, Puri P. Deficiency of platelet-derived
mimic Westphalen’s observations in Ret/ mice. Further growth factor receptor-α-positive cells in Hirschsprung’s disease
research will be necessary to elucidate the exact role of tuft colon. World J Gastroenterol 2016;22(12):3335–3340
cells in the pathophysiology of HSCR. 9 O’Donnell AM, Nakamura H, Tomuschat C, Marayati NF, Puri P.
DCLK1 and choline acetyltransferase (ChAT), the enzyme Altered expression of KCNG3 and KCNG4 in Hirschsprung’s dis-
ease. Pediatr Surg Int 2019;35(02):193–197
responsible for acetylcholine (ACh) production, are reported
10 O’Donnell AM, Nakamura H, Tomuschat C, Marayati NF, Puri P.
to be coexpressed within the gastrointestinal tract.1 Non- Abnormal Scn1b and Fxyd1 gene expression in the pulled-through
neurally produced ACh is an unusual concept. Schütz et al ganglionic colon may influence functional outcome in patients with
demonstrated that most tuft cells along the GI and biliary Hirschsprung’s disease. Pediatr Surg Int 2019;35(01):9–14

11 Cheng X, Voss U, Ekblad E. Tuft cells: distribution and connections 13 Cheng X, Voss U, Ekblad E. A novel serotonin-containing tuft cell
with nerves and endocrine cells in mouse intestine. Exp Cell Res subpopulation in mouse intestine. Cell Tissue Res 2019;376(02):
2018;369(01):105–111 189–197
12 Middelhoff M, Westphalen CB, Hayakawa Y, et al. Dclk1-express- 14 Westphalen CB, Asfaha S, Hayakawa Y, et al. Long-lived intestinal
ing tuft cells: critical modulators of the intestinal niche? Am J tuft cells serve as colon cancer-initiating cells. J Clin Invest 2014;
Physiol Gastrointest Liver Physiol 2017;313(04):G285–G299 124(03):1283–1295

Published online: 03.01.2020
Review Article
Management of Spontaneous Pneumothorax in

Children: A Systematic Review and Meta-Analysis
Maria Enrica Miscia1 Giuseppe Lauriti1 Gabriele Lisi1 Angela Riccio1 Pierluigi Lelli Chiesa1
1 Department of Pediatric Surgery, “Spirito Santo” Hospital, Pescara, Address for correspondence Giuseppe Lauriti, MD, PhD, Department
“G. d’Annunzio” University, Chieti-Pescara, Italy of Pediatric Surgery, “Spirito Santo” Hospital, Pescara, “G.
d’Annunzio” University, Chieti-Pescara, Via Fonte Romana 8, Pescara
Eur J Pediatr Surg 65100, Italy (e-mail: [email protected]).
Abstract Introduction Management of primary spontaneous pneumothorax (PSP) is mainly

based on adults. Data are controversial with regards to its management in children. We
aimed to assess: (1) the length of hospital stay (LOS) between conservative manage-
ment (i.e., observation with O2 administration), aspiration/chest drain, and surgical
management; (2) the risk of recurrence after nonsurgical treatment versus surgery; (3)
the risk of recurrence in the presence of bullae.
Materials and Methods Using a defined search strategy, three independent investigators
identified all the studies on the management of PSP in children. Case reports, opinion articles,
and gray literature publications were excluded. The study was conducted according to the
PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. A
meta-analysis was performed using RevMan 5.3. Data are expressed as mean SD.
Results Of 3,089 abstracts screened, 95 full-text were analyzed, 23 were included in
the quantitative analysis, and 16 were included in the meta-analysis (1,633 patients).
LOS was similar between conservative and surgical management (6.2 0.8 days vs.
5.9 1.4 days; p ¼ ns). Recurrence of PSP was significantly higher among children with
a nonsurgical management (32%) versus those surgically treated (18%; p ¼ 0.002). The
incidence of recurrence was slightly higher in patients managed by aspiration/chest
drain (34%) compared with those with a conservative management (27%; p ¼ 0.05).
Keywords Risk of recurrence in patients with or without documented bullae was not significantly
► spontaneous different (26 vs. 38%, respectively; p ¼ ns).
pneumothorax Conclusion Given the lack of a standardized management of pediatric PSP, the
► children present study seems to demonstrate a better outcome in children treated with surgery
► systematic review as first-line of management.
► meta-analysis Level of Evidence This is a Level III study.
Introduction The incidence of PSP in the pediatric population is low and

it has been reported to be around 3.4/100,000 children, with
Primary spontaneous pneumothorax (PSP) is defined as a a male to female ratio of 4:1 and a peak of incidence during
pneumothorax occurring neither with an underlying lung the adolescence.4,10–13 However, it has been reported a
disease nor a trauma.1–5 higher recurrence rate in the pediatric population compared
The incidence of PSP ranges between 4.7 and 28/100,000/ with adults (50–60% vs. 30–50%).5,7,10,12,14,15
year in men and 1.2 and 6/100,000/year in women, with a Air containing lesions (blebs or bullae) are thought to play
higher prevalence in young, tall, and thin males.1,2,4–9 an important role in the etiopathogenesis of PSP.1,7–9,15–17

November 21, 2019
Management of Spontaneous Pneumothorax in Children Miscia et al.
Guidelines exist from the British Thoracic Society and Table 1 Inclusion criteria of systematic review
American College of Chest Physicians regarding the manage-
ment of PSP in adults. However, there is not a consensus Publication
regarding the strategy of treatment of PSP in children up to Language Any
now.1,5,9,15,18 Date After 1950
Moreover, there are three reported methods for calculating
Subject Human studies
the size of pneumothorax on chest X-ray in adult population:
the Light, Rhea and Collins algorithms.8 The British Thoracic Study type Retrospective
Society guidelines suggest defining a pneumothorax as large if Prospective
there is a 2 cm gap between the lateral lung edge and the Case control
chest wall at the level of the hilum.9 The American College of
Cohort
Chest Physicians guidelines on the other side consider a large
pneumothorax when there is an apical distance 3 cm Excluded Case reports
between the thoracic wall and the lung.18 Case series
Up till now, the management of pediatric PSP is mainly Letters
based on the clinical conditions of the patients, with some
Editorials
authors suggesting an early surgical procedure due to the high
risk of recurrence.7,11,14 Conversely, others prefer an initial Gray literature
nonoperative treatment (with O2 administration, needle aspi- Keywords Primary spontaneous pneumothorax
ration and/or chest drain), reserving surgery in case of recur- Spontaneous pneumothorax
rence or persistent air leak.3,5,12,17,19
Children
Moreover, controversies exist on the role of the
Nonsurgical management
blebs/bullae as predicting factors for recurrence.7,12,15
With the present study we aimed to assess: (1) the length of Surgical management
hospital stay (LOS) between nonsurgical treatment (defined as
observation alone with O2 administration, needle aspiration
and/or chest drain) and surgical management; (2) the risk of included all studies (trials, cohort, and case-control) that
recurrence after nonsurgical treatment versus surgery; (3) the reported at least one outcome of interest. Furthermore, we
risk of recurrence in the presence of air-containing lesions included in the meta-analysis only those studies comparing
(blebs/bullae). different managements of PSP in children. If two or more
studies reported overlapping patient cohorts, for each out-
come measure we included only the article with the largest
number of patients. Any disagreement over the eligibility of a
Data Sources and Study Selection specific study was resolved through the discussion with the
This study was registered on PROSPERO—international pro- fourth author (G.Li). Moreover, the size of PSP of all included
spective register of systematic reviews (registration number: studies has been reported, when mentioned.
CRD42019131588).20 The systematic review was drafted The outcome measures for the present study were:
according to the PRISMA (Preferred Reporting Items for
1. The LOS between conservative management (i.e., obser-
Systematic Reviews and Meta-Analyses) statement.21
vation with O2 administration), aspiration and/or chest
A systematic review of the literature was performed
drain, and surgical management;
using a defined search strategy (►Table 1). Three investiga-
2. The risk of recurrence after conservative treatment versus
tors (M.E.M., G.La, and A.R.) independently screened
aspiration and/or chest drain versus surgery;
scientific databases (PubMed, Medline, Cochrane Collabora-
3. The risk of recurrence in the presence of air containing
tion, Scopus, and Ovid) looking for studies reporting
lesions (blebs/bullae) at computed tomography (CT) scan.
on spontaneous pneumothorax in children. Medical Sub-
ject Headings (MESH) and terms used are “spontaneous Statistical Analysis
pneumothorax,” “spontaneous pneumothorax AND manage- Categorical variable frequencies were compared using Pear-
ment,” and ‘‘spontaneous pneumothorax AND children” son’s chi-square test or the two-tailed Fisher exact probability
(►Supplementary Table S1, available in the online version). test, as appropriate. When median and range were reported,
Reference lists were examined to identify relevant cross- mean SDs were estimated, as previously reported.22 Meta-
references. Case reports, opinion articles, experimental analysis of comparative studies was conducted using RevMan
studies, and case series with less than 10 patients were 5.3.23 Data are presented as risk ratio (RR) for categorical
excluded. All gray literature publications (i.e., reports, theses, variables, and mean differences (MD) for continuous variables,
conference proceedings, bibliographies, commercial docu- along with 95% confidence intervals (CI) using the random-
mentations, and official documents not published commer- effects model, with p-values shown for Z-test for overall
cially) were excluded. Full text articles of potentially eligible significance and I2 statistic for heterogeneity. A p-value
studies were retrieved and independently assessed for suit- <0.05 was considered statistically significant. Data are
ability by three investigators (M.E.M., G.La, and A.R.). We expressed as mean SD.

Quality Assessment 100% were considered as low, moderate, substantial, and of

Risk of bias for individual studies was assessed in duplicate (M.E. considerable heterogeneity, respectively. Imprecision was
M. and A.R.) using the methodological index for nonrandomized assessed using optimal information size, which was based on
studies (MINORS).24 Differences between the two reviewers (M. 25% relative risk reduction, 0.05 of α error, and 0.20 of β error.26
E.M. and A.R.) were resolved through consensus and discussion
with the third author (G.La). The total score for this 12-item
Results
instrument ranges 0 to 24 points with a validated “gold stan-
dard” cut-off of 19.8. We assessed the methodological quality for Study Selection and Characteristics
each outcome by grading the quality of evidence using the Of 3,049 abstracts screened, 95 full-text were analyzed, 23
Grading of Recommendations Assessment, Development and were included in the quantitative analysis,2–7,10–17,19,27–34
Evaluation (GRADE) methodology.25 Quality of evidence was and 16 were included in the meta-analysis (1,633
rated as high, moderate, low, and very low for each outcome. patients, ►Fig. 1, ►Table 2).2,3,6,7,11–15,17,19,27,28,31,33,34
Observational studies start with a low quality of evidence. The The LOS was similar between nonsurgical and surgical
quality of evidence was rated down in the presence of risk of management of PSP (5.9 1.4 vs. 6.2 0.8 days, respectively;
bias, inconsistency, indirectness, imprecision, and publication p ¼ ns, MD 0.27 [95% CI 0.11, 0.64], I2 ¼ 0%; ►Fig. 2A).
bias. For assessment of risk of bias in observational studies, we Similarly, the LOS was similar between conservative man-
used the MINORS instrument. Inconsistency was determined agement of PSP compared with aspiration and/or chest drain
according to heterogeneity. We produced I2 values to assess (4.7 1.2 vs. 7.3 1.5 days, respectively; p ¼ ns, MD 1.41
heterogeneity. I2 value of 0 to 40, 30 to 60, 50 to 90, and 75 to [95% CI 0.78, 3.60], I2 ¼ 97%; ►Fig. 2B).
Fig. 1 Diagram of workflow in the systematic review and meta-analysis.

Table 2 Studies included in the meta-analysis
Author Year Type of LOS (days) Surgery during Recurrence Recurrence (bullae)
study 1st admission

O2 A/CD Surg. O2 (%) A/CD (%) O2 (%) A/CD (%) Surg. (%) Bullae (%) No bullae (%)
34
Yu et al 1975 R n.r. n.r. n.r. n.r. n.r. 0/8(0) 0/2(0) n.r. n.r. n.r.
33
Davis et al 1993 R n.r. n.r. n.r. n.r. n.r. 0/1(0) 5/5(100) 1/6(17) n.r. n.r.
Poenaru et al14 1994 R 2.9 5.5 n.r. n.r. n.r. 3/23(13) 7/35(20) n.r. n.r. n.r.
Wilcox et al17 1995 R n.r. n.r. n.r. 0/3(0) 4/14(29) 0/3(0) 3/14(21) n.r. n.r. n.r.
19
Qureshi et al 2005 R n.r. n.r. n.r. n.r. n.r. 20/37(54) 4/14(29) n.r. n.r.
13
Chung et al 2009 R n.r. n.r. n.r. n.r. 9/15(60) n.r. 0/15(0) n.r. 0/7(0) 0/8(0)
Nathan et al6 2010 R n.r. n.r. n.r. n.r. n.r. 6/25(24) 3/14(21) 3/11(27)
3
Lee et al 2010 R n.r. n.r. n.r. n.r. n.r. 4/12(33) 30/59(51) 0/6(0) 1/26(4) 3/8(37)
2
Zganjer et al 2010 R n.r. n.r. n.r. n.r. n.r. n.r. 0/2(0) 1/14(7) 1/12(8) 0/4(0)
Management of Spontaneous Pneumothorax in Children
Seguier-Lipszyc 2011 R n.r. n.r. n.r. n.r. n.r. 10/18(55) 8/18(44) 0/10(0) 5/13(38) 15/33(45)
et al7
Smith et al31 2011 R 4.9(3–25)a 11.7(5–32)a n.r. n.r. n.r. n.r. n.r. n.r. n.r. n.r.
12
Lopez et al 2014 R n.r. n.r. n.r. 8/25(32) 28/73(38) 10/25(40) 25/73(34) 13/89(15) n.r. n.r.
Miscia et al.
Young Choi 2014 R 4.5 1 3.6 1 n.r. n.r. n.r. 18/42(43) 36/72(50) n.r. 12/14(80) 42/100(42)
et al15
Soler et al11 2018 R n.r. n.r. n.r. n.r. n.r. 11/33(33) 19/33(58) 2/14(14) n.r. n.r.
28
Williams et al 2018 R 4.1 4 7.2 7 6.2 4 53/336(16) 157/497 (32) 140/497 (28) 82/366 (22) 42/207(20) n.r. n.r.
27 b b b
Williams et al 2018 R 5.5(3–7) 6 (3–7.5) 5.5(5–7) 5/8(62) 14/28(50) 3/8(37) 10/28(34) 4/10(40) n.r. n.r.
Abbreviations: a, median (range); b, median (IQR); A/CD, aspiration and/or chest drain; n.r., not reported; O2, conservative management with oxygen administration; P, prospective cohort study; PD, prospective
database; R, retrospective; Surg., surgery.
Note: Data are expressed as mean SD.
Fig. 2 (A) Forest plot comparison of the length of hospital stay (LOS) between nonsurgical and surgical management of primary spontaneous
pneumothorax (PSP). (B) Forest plot comparison of the LOS between conservative management and aspiration and/or chest drain (A/CD) of PSP.
Recurrence of PSP was significantly more frequent among surgical procedure during the first hospital admission was
children with a nonsurgical management (379/1,185 patients, similar in those managed by aspiration and/or chest drain
31.9%) versus those surgically treated (67/370 cases, 18.1%; (203/612 patients, 33.2%) compared with those with a conser-
p ¼ 0.002, RR 0.52 [95% CI 0.35, 0.79], I2 ¼ 37%; ►Fig. 3A). vative management (66/372 cases, 17.7%; p ¼ ns, RR 1.36 [95%
Moreover, the incidence of recurrence was higher albeit not CI 0.79, 2.33], I2 ¼ 62%; ►Fig. 4).
significant in those managed by aspiration and/or chest drain The risk of recurrence in children with documented bullae
(283/836 patients, 33.8%) compared with those with a conser- at CT scan (22/86 patients, 26%) was not significantly different
vative management (139/509 cases, 27.3%; p ¼ 0.05, RR 1.18 compared with those with no bullae detected (63/164 cases,
[95% CI 1.00, 1.40], I2 ¼ 0%; ►Fig. 3B). However, the risk of a 38%; p ¼ ns, RR 0.84 [95% CI 0.30, 2.36], I2 ¼ 78%; ►Fig. 5).
Fig. 3 (A) Forest plot comparison of recurrence of primary spontaneous pneumothorax (PSP) among children with a nonsurgical management
versus those surgically treated. (B) Forest plot comparison of recurrence of PSP among children managed by aspiration and/or chest drain (A/CD)
compared with those with a conservative management.

Fig. 4 Forest plot comparison of the risk of surgical procedure during the first hospital admission between children managed by aspiration
and/or chest drain (A/CD) compared with those with a conservative management.
Fig. 5 Forest plot comparison of the risk of recurrence in children with documented bullae at computed tomography scan compared with those
with no bullae detected.
Size of Pneumothorax The higher incidence of PSP among neonates (mainly

Only 5 out of 16 studies included in the meta-analysis those preterms) may be explained by the high transpulmo-
analyzed the size of the PSP (►Table 3). Moreover, different nary pressure generated at the first breath,31 while
methods have been used to determine this measure. the second peak of incidence during adolescence has been
connected to the rupture of apical bullae/blebs.1,7–9,15–17
Moreover, even if it has been reported a higher incidence
Discussion
of PSP in adults, this has not been confirmed with regards to
The correct management of pediatric PSP has always been the recurrence rate. As a matter of fact, the incidence of
object of controversies. recurrence has been reported to be around 30% in adults and
Several studies have been published on this topic up to now; up to 50 to 60% in children.5,12,15,35
however, they are mainly retrospective with a limited cohort of On this basis, questions have raised whether children
patients or case series.2–8,10,12–17,19,27–34 Moreover, data from should undergo a more aggressive initial treatment or not.10,19
children are often included in larger adult series, thus resulting Historically, the first-line of management of PSP in children
difficult to be deduced. has been conservative, since traditional thoracotomy was
Therefore, pediatric guidelines are still lacking and children more invasive and painful. Moreover, thoracotomy usually
with PSP are commonly managed according to the experience required a lengthened hospital stay.12,17,33 However, the
and preference of different clinicians.5,8 advent of the video-assisted thoracoscopic surgery (VATS),
Possible managements consist of observation with O2 guaranteed a less invasive surgical approach with a shortened
administration, needle aspiration, chest drain insertion, or hospital stay. These improvements produced a shift toward an
thoracotomic/thoracoscopic surgery (blebectomy/apicectomy earlier surgical management to prevent recurrences.12
with pleurodesis).9 Controversies exist also in terms of quantification of pneu-
The incidence of pediatric PSP is lower than the one mothorax. As mentioned, few methods for calculating the size
reported in adults and it is estimated around 3.4/100,000 of PSP have been reported.8,9,18 However, all of them are
children. Boys are more commonly affected than girls, with complexes and not very accurate.8 Based on the different
a male:female ratio of 4:1.4,10–13 However, Poenaru et al in guidelines used, the management of the pneumothorax can
1994 showed that boys and girls under 9 years of age are differ. Moreover, both the British Thoracic Society and the
equally affected.14 American College of Chest Physicians guidelines enlighten
The incidence of PSP among the pediatric population has a the importance of the clinical features of the patients more
bi-phasic distribution: there is a first peak of incidence in the than the size of the pneumothorax in the management of
neonatal period and a second one throughout adolescence.13 the PSP. In fact, a clinically stable large pneumothorax can be

Table 3 Methods of quantification of pneumothorax
Author Year Pneumothorax quantification

34
Yu et al 1975 N/A
Davis et al33 1993 N/A
4
Poenaru et al 1994 N/A
17
Wilcox et al 1995 N/A
Qureshi et al19 2005 N/A
13
Chung et al 2009 N/A
6
Nathan et al 2010 Degree 1: partial slide of pneumothorax
Degree 2: complete pneumothorax
Degree 3: complete pneumothorax with contralateral shift of the mediastinum
Lee et al3 2010 Light index; Collins formula
2
Zganjer et al 2010 Small: rim of air <2 cm
Moderate: lung collapsed halfway to the heart border
Large: lung completely collapsed
Seguier-Lipszyc et al7 2011 N/A
31
Smith et al 2011 N/A
Lopez et al12 2014 N/A
15
Young Choi et al 2014 Distance from the chest wall to the pleural line:
Minimal: <1 cm
Small: between 1 and 2 cm
Moderate: distance between small and large
Large: distance > width of the remaining lung
Complete: when the lung mass was restricted to hilum
Soler et al11 2018 Small: <2 ribs visible on chest-X ray or <15%
Moderate: 2–4 ribs or 15–30%
Large: >4 ribs or >30%
Williams et al27 2018 N/A
28
Williams et al 2018 N/A
managed conservatively while a patient with dyspnea and chest tube insertion.15 Also, Lee et al inserted a chest tube
chest pain, even if with a small pneumothorax should undergo unless the pneumothorax was “just a small rim of air around
aspiration and/or chest drain.9,18 Furthermore, the manage- the lung.”3
ment of the PSP mainly depended on the presence of symptoms Furthermore, a retrospective multicenter study by Soccorso
more than on the size of the pneumothorax: asymptomatic et al in 2015 based on 50 pediatric cases showed that 53% of
patients were always treated with a nonoperative management children treated by needle aspiration recurred and ultimately
(i.e., observation alone and/or O2 administration). required a chest drain. Thus, the authors suggested to directly
Based on the current adults guidelines, clinically stable insert a chest tube in children to reduce the risk of repeat
patients with PSP should be treated conservatively, with procedures and anesthesia.29An early VATS has also been
observation alone or O2 administration, thus reserving active proposed to avoid the risk of recurrence and thus to reduce
intervention for breathless patients. Even if British Thoracic the total hospital stay in children who should undergo a chest
Society guidelines proposed needle aspiration as first-line tube insertion for the first episode of PSP.7,27,28
management for symptomatic PSP, American College of Williams et al reported that the length of stay was similar
Chest Physicians suggested the use of chest drain because between patients treated by surgery or aspiration and/or
of the high risk of additional procedures and persistent air chest drain at their first episode.27 The results of our meta-
leaking following the needle aspiration.9,18 analysis seem to corroborate this outcome.
Commonly, when symptoms such as respiratory distress Moreover, it has reported a recurrence rate after the
occurred, needle aspiration or chest tube insertion was primary surgery between 4 and 40%.7,12,19,27,28 These data,
performed.2,6,11 In case of a persistent air leaking, children even if low, seem not justify an initial surgical approach.
were candidate for surgery.2,3,6,11,15 In our study we found a significantly lower risk of recur-
Nonetheless, Young Choi et al suggested that asymptomatic rence between patients treated by primary surgery versus
patients with small pneumothorax should be managed with those nonsurgically managed (18 vs. 32%, respectively;
O2 administration, while small, symptomatic pneumothorax p ¼ 0.002) and among the latter, those managed by observa-
or pneumothorax of any other size should be managed with tion alone had a slightly reduced risk of recurrence than those

Table 4 Risk of bias assessment for individual studies using methodological index for nonrandomized studies (MINORS)24
Item Yu Davis Poenaru Wilcox Qureshi Chung Nathan Lee Zganjer Seguier- Smith Lopez Young Soler Williams Williams
et al34 et al33 et al14 et al17 et al19 et al13 et al6 et al3 et al2 Lipszyc et al31 et al12 Choi et al11 et al28 et al27
et al7 et al15

1. A clearly stated aim 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
2. Inclusion of consecutive 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
patients
3. Prospective collection of data 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
4. End points appropriate to the 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
aim of the study
5. Unbiased assessment of 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
the study end point
6. Follow-up period appropriate 1 2 0 1 0 0 2 2 2 2 0 0 1 0 0 0
to the aim of the study
7. Loss to follow-up 0 1 0 0 0 0 1 2 0 0 0 0 1 0 0 0
less than 5%
8. Prospective calculation of 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
the study size
9. An adequate control group 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
10. Contemporary groups 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
Miscia et al.
11. Baseline equivalence 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2

of groups
12. Adequate statistical 0 0 0 0 2 0 2 2 2 2 2 2 2 2 2 2
analyses
Total score 13 15 12 15 14 12 17 18 16 16 14 14 16 14 14 14
0 ¼ not reported.
1 ¼ reported but inadequate.
2 ¼ reported and adequate.
Table 5 GRADE evidence profile25 for the management of PSP in children
Quality assessment No. of patients Effect Quality

No. of Study Risk Inconsistency Indirectness Imprecision Other Cases Controls Relative Absolute
studies design of bias considerations (95% CI) (95% CI)
LOS in surgery vs. no surgery Surgery No surgery
2 OS Moderatea Substantial Not serious Seriousb None 217 869 – MD 0.27 higher ⊗OOO
(0.12 lower to VERY LOW
0.71 higher)
LOS in A/CD vs. conservative A/CD Conservative
a b
4 OS Moderate Considerable Not serious Serious None 547 398 – MD 1.41 higher ⊗⊗OO
(1.44 lower to LOW
6.64 higher)
Recurrence in surgery vs. no surgery Surgery No surgery
a b
9 OS Moderate Low Not serious Serious None 67/370 379/1,185 RR 0.52 138 fewer per 1,000 ⊗⊗OO
(18.1%) (31.9%) (0.35, 0.79) (from 93 fewer to LOW
210 fewer)
Recurrence in A/CD vs. conservative A/CD Conservative
a b
11 OS Moderate Low Not serious Serious None 283/836 139/509 RR 1.18 65 more per 1,000 ⊗⊗OO
(33.8%) (27.3%) (1.00, 1.40) (from 55 more to LOW
77 more)
Surgery at first admission in A/CD vs. conservative A/CD Conservative
a b
4 OS Moderate Considerable Not serious Serious None 203/612 66/372 RR 1.36 ⊗⊗OO
(33.2%) (17.7%) (0.79, 2.33) LOW
Recurrence in bullae vs. no bullae Bullae No bullae
a b
6 OS Moderate Considerable Not serious Serious None 22/86 63/164 RR 0.84 128 fewer per 1000 ⊗⊗OO
(25.6%) (38.4%) (0.30, 2.36) (from 46 fewer to LOW
360 more)
Abbreviations: A/CD, aspiration and/or chest drain; CI, confidence interval; LOS, length of hospital stay; MD, mean deviation; OS, observational study; PSP, primary spontaneous pneumothorax; RR, risk relative.
a
Bias due to possible confounding.
b
OIS, optimal information size not met
Note: GRADE working group grades of evidence.
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

Miscia et al.

who underwent aspiration and/or chest drain (27 vs. 34%, Conclusion
respectively; p ¼ 0.05).
The reduced number of procedures can explain the This is the first systematic review and meta-analysis com-
shorter hospital stay in those primarily treated by surgery, paring the outcomes between conservative management,
thus eliminating the time-lapse between the first nonsurgi- aspiration and/or chest drain, and surgical procedure of PSP
cal management and the surgical one. in children. Given the lack of high-quality studies and the
Moreover, the risk of a surgical procedure did not differ paucity of information related to the size of the pneumotho-
between those children initially managed either by conser- rax, primary surgery seems to reduce the risk of recurrence.
vative or operative treatment. This might support the choice However, a surgical procedure could be safely performed as
of a primary conservative management followed by an early early as possible either after a primary conservative
surgery if there is a persistent air leak (>3–5 days) or a management or aspiration and/or chest drain, without an
recurrence, in clinically stable patients.8,35 increased LOS.
The role of air-containing lesions (blebs/bullae) in the etio- The presence of bullae seems not to predict a higher risk of
pathogenesis and in the risk of recurrence of PSP has also been recurrence, thus not supporting the need of a high-resolu-
debated, along with the need of performing a high-resolution tion CT to all pediatric patients after the first episode of PSP.
CT in children with a first episode of PSP. It has been reported an However, further data given by high-quality studies
incidence of air-containing lesions at high-resolution CT rang- would be needed to corroborate our preliminary results.
ing from 31 to 100% in children with PSP. Moreover, the
recurrence rate of pneumothorax in these children has been Conflict of Interest
reported to range between 50 and 100%, thus suggesting an None declared.
important role of the blebs in the risk of recurrence.7,15
A retrospective single center study on 114 children from
Young Choi et al showed a 55% incidence of bullae on high- References
resolution CT in patients with a first episode of PSP with a 1 Brown SG, Ball EL, Perrin K, et al. Study protocol for a randomised
risk of recurrence in this cohort of 60%.15 Soccorso et al, in controlled trial of invasive versus conservative management of
primary spontaneous pneumothorax. BMJ Open 2016;6(09):
2015, reported a 100% incidence of bullae in the surgical
e011826
specimens of children who underwent wedge resection after 2 Zganjer M, Cizmić A, Pajić A, Cigit I, Zganjer V. Primary spontane-
a first episode of PSP.29 On the other side, some authors do ous pneumothorax in pediatric patients: our 7-year experience.
not recommend the use of high-resolution CT for all the J Laparoendosc Adv Surg Tech A 2010;20(02):195–198
children with PSP due to its low sensitivity as screening test 3 Lee LP, Lai MH, Chiu WK, Leung MW, Liu KK, Chan HB. Manage-
and its high radiation dose.8,27 ment of primary spontaneous pneumothorax in Chinese children.
Hong Kong Med J 2010;16(02):94–100
In our study we reported that the presence of bullae did
4 Dotson K, Timm N, Gittelman M. Is spontaneous pneumothorax
not increase the risk of recurrence of PSP, thus supporting the really a pediatric problem? A national perspective. Pediatr Emerg
hypothesis that a high-resolution CT seems not to be man- Care 2012;28(04):340–344
datory after the first episode of PSP. 5 Matuszczak E, Dębek W, Hermanowicz A, Tylicka M. Spontaneous
pneumothorax in children—management, results, and review
Limitations of the literature. Kardiochir Torakochirurgia Pol 2015;12(04):
322–327
We are aware of the limitations of our study, which rely on the
6 Nathan N, Guilbert J, Larroquet M, Lenoir M, Clement A, Epaud R.
quality of the studies and data available in the literature, as any Efficacy of blebs detection for preventive surgery in children’s
other meta-analysis. Moreover, we highlighted the lack of idiopathic spontaneous pneumothorax. World J Surg 2010;34
consensus in the definition of the size of pneumothorax. (01):185–189
Therefore, a limitation of our study is the inability to measure 7 Seguier-Lipszyc E, Elizur A, Klin B, Vaiman M, Lotan G. Manage-
ment of primary spontaneous pneumothorax in children. Clin
the effect of the size of the PSP. Moreover, all the studies
Pediatr (Phila) 2011;50(09):797–802
included in the meta-analysis were retrospective and only
8 Robinson PD, Cooper P, Ranganathan SC. Evidence-based manage-
few of them have reported a follow-up period appropriate to ment of paediatric primary spontaneous pneumothorax. Paediatr
the aim of the study. None of the papers provided sample size Respir Rev 2009;10(03):110–117
calculations, and most of the studies lack in few of the outcomes 9 MacDuff A, Arnold A, Harvey J; BTS Pleural Disease Guideline
considered (►Table 2). As a consequence, none of the studies Group. Management of spontaneous pneumothorax: British Tho-
racic Society Pleural Disease Guideline 2010. Thorax 2010;65
included reached the gold standard cut-off on MINORS of 19.8
(Suppl 2):ii18–ii31
out of 24 (►Table 4). According to the GRADE methodology, the 10 Cook CH, Melvin WS, Groner JI, Allen E, King DR. A cost-effective
quality of evidence was low for the reported outcomes thoracoscopic treatment strategy for pediatric spontaneous
(►Table 5). However, when independently assessed by two pneumothorax. Surg Endosc 1999;13(12):1208–1210
authors (G.Li and P.L.C.) using AMSTAR (A Measurement Toll to 11 Soler LM, Raymond SL, Larson SD, Taylor JA, Islam S. Initial
Assess Systematic Reviews),36 the present systematic review primary spontaneous pneumothorax in children and adolescents:
operate or wait? J Pediatr Surg 2018;53(10):1960–1963
and meta-analysis received a relevant score (►Supplementary
12 Lopez ME, Fallon SC, Lee TC, Rodriguez JR, Brandt ML, Mazziotti
Table S2, available in the online version). The PRISMA checklist MV. Management of the pediatric spontaneous pneumothorax: is
was then completed (►Supplementary Table S3, available in primary surgery the treatment of choice? Am J Surg 2014;208
the online version). (04):571–576

13 Chung PH, Wong KK, Lan LC, Tam PK. Thoracoscopic bullectomy 25 Guyatt GH, Oxman AD, Vist GE, et al; GRADE Working Group.
for primary spontaneous pneumothorax in pediatric patients. GRADE: an emerging consensus on rating quality of evidence and
Pediatr Surg Int 2009;25(09):763–766 strength of recommendations. BMJ 2008;336(7650):924–926
14 Poenaru D, Yazbeck S, Murphy S. Primary spontaneous pneumo- 26 Dupont WD, Plummer WD Jr. Power and sample size calculations.
thorax in children. J Pediatr Surg 1994;29(09):1183–1185 A review and computer program. Control Clin Trials 1990;11(02):
15 Young Choi S, Beom Park C, Wha Song S, et al. What factors predict 116–128
recurrence after an initial episode of primary spontaneous pneu- 27 Williams K, Lautz TB, Leon AH, Oyetunji TA. Optimal timing of
mothorax in children? Ann Thorac Cardiovasc Surg 2014;20(06): video-assisted thoracoscopic surgery for primary spontaneous
961–967 pneumothorax in children. J Pediatr Surg 2018;53(09):1858–1861
16 Beg MH, Reyazuddin, Faridi MM, Ahmad SH, Shahab T. Spontane- 28 Williams K, Oyetunji TA, Hsuing G, Hendrickson RJ, Lautz TB.
ous pneumothorax in children—a review of 95 cases. Ann Trop Spontaneous pneumothorax in children: national management
Paediatr 1988;8(01):18–21 strategies and outcomes. J Laparoendosc Adv Surg Tech A 2018;28
17 Wilcox DT, Glick PL, Karamanoukian HL, Allen JE, Azizkhan RG. (02):218–222
Spontaneous pneumothorax: a single-institution, 12-year expe- 29 Soccorso G, Anbarasan R, Singh M, Lindley RM, Marven SS, Parikh
rience in patients under 16 years of age. J Pediatr Surg 1995;30 DH. Management of large primary spontaneous pneumothorax in
(10):1452–1454 children: radiological guidance, surgical intervention and pro-
18 Baumann MH, Strange C, Heffner JE, et al; AACP Pneumothorax posed guideline. Pediatr Surg Int 2015;31(12):1139–1144
Consensus Group. Management of spontaneous pneumothorax: 30 Noh D, Lee S, Haam SJ, Paik HC, Lee DY. Recurrence of primary
an American College of Chest Physicians Delphi consensus state- spontaneous pneumothorax in young adults and children. Inter-
ment. Chest 2001;119(02):590–602 act Cardiovasc Thorac Surg 2015;21(02):195–199
19 Qureshi FG, Sandulache VC, Richardson W, Ergun O, Ford HR, 31 Smith J, Schumacher RE, Donn SM, Sarkar S. Clinical course of
Hackam DJ. Primary vs delayed surgery for spontaneous pneu- symptomatic spontaneous pneumothorax in term and late pre-
mothorax in children: which is better? J Pediatr Surg 2005;40 term newborns: report from a large cohort. Am J Perinatol 2011;
(01):166–169 28(02):163–168
20 PROSPERO international prospective register of systematic reviews. 32 Chambers A, Scarci M. In patients with first-episode primary
Available at: https://www.crd.york.ac.uk/prospero. Accessed May 15, spontaneous pneumothorax is video-assisted thoracoscopic sur-
2019 gery superior to tube thoracostomy alone in terms of time to
21 Moher D, Liberati A, Tetzlaff J, Altman DG; PRISMA Group. Preferred resolution of pneumothorax and incidence of recurrence? Inter-
reporting items for systematic reviews and meta-analyses: the act Cardiovasc Thorac Surg 2009;9(06):1003–1008
PRISMA statement. J Clin Epidemiol 2009;62(10):1006–1012 33 Davis AM, Wensley DF, Phelan PD. Spontaneous pneumothorax in
22 Hozo SP, Djulbegovic B, Hozo I. Estimating the mean and variance paediatric patients. Respir Med 1993;87(07):531–534
from the median, range, and the size of a sample. BMC Med Res 34 Yu VY, Liew SW, Robertson NR. Pneumothorax in the newborn.
Methodol 2005;5:13. doi:10.1186/1471-2288-5-13 Changing pattern. Arch Dis Child 1975;50(06):449–453
23 Review Manager 5 (RevMan). The Cochrane Collaboration. Copen- 35 Williams K, Baumann L, Grabowski J, Lautz TB. Current practice in the
hagen: The Nordic Cochrane Centre; 2014 management of spontaneous pneumothorax in children. J Laparoen-
24 Slim K, Nini E, Forestier D, Kwiatkowski F, Panis Y, Chipponi J. dosc Adv Surg Tech A 2019;29(04):551–556
Methodological index for non-randomized studies (minors): 36 Shea BJ, Grimshaw JM, Wells GA, et al. Development of AMSTAR: a
development and validation of a new instrument. ANZ J Surg measurement tool to assess the methodological quality of sys-
2003;73(09):712–716 tematic reviews. BMC Med Res Methodol 2007;7:10

Original Article
Outcomes of Primary versus Multiple-Staged

Repair in Hirschsprung’s Disease in England
Stefano Giuliani1, Kate Honeyford2, Chieh-Yu Chang2 Alex Bottle2 Paul Aylin2
1 Division of Specialist Neonatal and Paediatric Surgery, Great Address for correspondence Stefano Giuliani, FRCS, MD, PhD,
Ormond Street Hospital for Children NHS Foundation Trust, London, Department of Neonatal and Paediatric Surgery, Great Ormond Street
United Kingdom Hospital for Children NHS Foundation Trust, 5th Floor Paul O’Gorman
2 Dr Foster Unit, Imperial College London, London, United Kingdom Building, Great Ormond Street, London WC1N 3JH, United Kingdom
(e-mail: [email protected]).
Eur J Pediatr Surg
Abstract Introduction The study aimed to compare 1-year outcomes for primary versus multiple-
staged (three operations with colostomy) repairs in Hirschsprung’s disease (HD).
Materials and Methods Retrospective analysis of a large national administrative
database (Hospital Episode Statistics) including all the neonates born with HD in
England between 2003 and 2015. Main outcomes were: 1-year mortality, postopera-
tive readmissions, and reoperations. Secondary outcomes: cumulative length of
hospital stay (cLOS) and hospital volume–outcome relationship.
Results A total of 1,333 neonates with HD were treated in 21 specialist pediatric
surgical centers; 874 (65.5%) patients had a primary repair for HD. One-year mortality
was 2.8%. The overall readmission rate was 70.2%, with a significant difference between
primary and multiple-staged repair (79.9 vs. 90.1%, p < 0.01). There was no difference
in reoperation. Primary pull-through was associated with a significantly lower proba-
Keywords bility of postoperative readmission (odds ratio [OR] ¼ 0.08, 95% confidence interval
► Hirschsprung’s [CI] ¼ 0.06–0.11, p < 0.001) and cLOS (OR ¼ 0.38, 95% CI ¼ 0.28–0.52, p < 0.001)
disease compared with multiple-staged repair. There were no significant difference in out-
► outcomes comes between patients treated in low-volume (<37 cases/year) and high-volume (>
► primary repair 55 cases/year) specialist centers.
► multiple-staged Conclusion Whenever clinically indicated, primary repair should be used in HD as this
► length of stay is associated with fewer readmissions and shorter time spent in the hospital.
Introduction
effective in cleaning the blocked stools. There are different
Hirschsprung’s disease (HD) is characterized by the absence surgical techniques for the definitive repair of this condition
of ganglion cells in the distal intestine leading to functional (i.e., Swenson, Soave, and Duhamel), and they can all be
intestinal obstruction in the neonatal period.1,2 The overall performed as a primary (one definitive surgery without
incidence of HD worldwide was estimated to be 2 in 10,000 colostomy) or multiple-staged (neonatal colostomy forma-
live births, and in the United Kingdom, it was 1.8 per 10,000 tion, definitive repair, and colostomy closure) repair.4
live births (150 new cases per year).3 Based on the length of Several reports have shown the feasibility and good out-
the colon affected by HD, a colostomy formation may be comes associated with primary repair in HD.5 At present, there
needed in the neonatal period if rectal washouts are not is no international consensus supporting the use of primary
versus multiple-staged repair in neonatal colorectal surgery.

Stefano Giuliani and Kate Honeyford share the first authorship. Significant variation has been highlighted by a recent

December 3, 2019
Outcomes of Primary versus Multiple-Staged Repair in HD in England Giuliani et al.
international survey in term of surgical technique used to or secondary pulmonary hypertension (I27.0, I27.2); chro-
repair HD.6 Surgeons’ preference for primary versus multiple- mosomal abnormalities (Q90–93, Q96–Q99, excludes Q93.6,
staged repair is often based on training received, level of Q95); and hypoplasia and dysplasia of lung (Q33.6).
experience, and attitude to adopt new surgical techniques.7,8
It is unknown if a primary repair can protect the neonates
Main Outcomes
from unnecessary multiple operations, higher complication
rate, or longer hospital stay. There is a lack of benchmark data Mortality, Readmissions, and Reoperations
in neonatal colorectal surgery on 1-year postoperative out- Overall 1-year mortality (using linked Office for National
comes and volume–outcome relationship among specialist Statistics death certificate data) was considered for all the
pediatric surgical centers (SPSC). patients operated in specialist centers. Overall readmissions
The study aimed to investigate the differences in 1-year within 1-year and 30 days were collected. Thirty-day emer-
postoperative outcomes between patients who underwent gency readmissions were defined as an admission to any NHS
primary versus multiple-staged repair for HD between 2003 hospital within 30 days after discharge from a SPSC. A wide
and 2015 in England. In addition, we studied the association range of OPCS-4 codes was used to identify unplanned oper-
between hospital volume and outcomes in specialist pediat- ations (reoperation) and these are listed in ►Supplementary
ric surgical centers. Appendix C (available in the online version). Patients who
underwent the index procedures plus any one of the
unplanned operations within 1 year were specified as
experiencing a reoperation (1-year cumulative reoperation).
Database Sources and Inclusion Criteria
The Hospital Episode Statistics (HES) database contains data Cumulative Length of Hospital Stay and Volume–
for patients admitted to all NHS hospital trusts in England, Outcome Relationship
collecting details of all admissions, outpatient appointments, Cumulative length of hospital stay (cLOS) was calculated as
and accident and emergency attendances.9,10 Each HES the sum of all the days spent in a SPSC within the first year of
record consists of information about how a patient is admit- life from admission to this hospital type. Hospital volumes
ted to an NHS hospital, including diagnostic, procedural, were based on the total case load for HD treated in each SPSC
demographic, administrative, and geographical informa- during the study years. In line with a methodology used in
tion.10 There are one primary diagnosis and 19 secondary other similar studies, the 21 specialist pediatric surgical
diagnoses in each episode, which are classified using the centers were apportioned into tertiles in line with the total
International Classification of Diseases, 10th revision (ICD- number of patients with procedures performed during the
10). Up to 24 procedures are also coded, using the Office of study period. The specialist hospitals were divided based on
Population Censuses and Surveys Classification of Interven- their volume of procedures per year: low- (<51/year), medi-
tions and Procedures, fourth version (OPCS-4). The Carstairs um- (51–68/year), and high-volume (>68/year) centers.13,14
index is a common index describing socioeconomic depriva-
tion, derived from the patients’ postcode and ethnic group, Statistical Analysis
which is contained in the HES record 11,12. Descriptive analysis was completed for patients with HD.
In our HES extract, we included all neonates born in England Chi-square test was used to assess associations between type
between April 1, 2003, and March 31, 2015, with an ICD-10 of operation (primary vs. multiple staged), patient and
diagnostic code of HD (aganglionosis, congenital [aganglionic] hospital characteristics, and patient outcomes.
megacolon, Q43.1) made within the first 28 days of life. Logistic regression was used to assess the associations
Records of patients admitted to the 21 specialist pediatric between procedure type and readmissions adjusting for
surgical centers (►Supplementary Appendix A, available in patient characteristics. All the models were adjusted by
the online version) were extracted and analyzed. age at admission, gender, ethnic group, socioeconomic dep-
Operative codes used to select primary and staged repairs rivation, year of admission, and comorbidities, using an α
are shown in ►Supplementary Appendix B (available in the level of 0.05. All the analyses were implemented in SAS 9.4.
online version). Postoperative outcomes were calculated
after the pull-through repair in the primary repair for HD.
Results
In case of multiple-staged repair, we considered reoperation
and readmission as the cumulative unplanned reoperations/ Patient Information and Trends
readmission found between the elective stages. Among 8,504,815 live births in English hospital data
We examined the most common comorbidities ICD-10 recorded between 2003 and 2015, 1,551 cases of HD were
codes associated with HD patients: extremely low birth diagnosed within their first 28 days of life (incidence of 1.82
weight (P07.0—newborn birth weight 999 g or less); other per 10,000 live births). Over the 12 years, the number of
low birth weight (P07.1—newborn birth weight 1,000– diagnosed cases remained steady, and boys accounted for
2,499 g); extreme immaturity (P07.2—less than 28 complet- 73.7% of all cases (n ¼ 1,143). The 365-day all-cause mortality
ed weeks of gestation); other preterm infants (P07.3—28 rate was 3.2% (50/1,551).
completed weeks or more but less than 37 completed weeks We identified 1,333 neonates (85.9% of all diagnosed cases
of gestation); major cardiac anomalies (Q20–Q26); primary in England) treated in the 21 specialist pediatric surgical

centers (►Table 1). There were minor fluctuations in the Operation Rate for Hirschsprung’s Disease
number of cases in the specialist centers during the study The majority of operations occurred in the specialist centers
period, but no overall trends were detected (cases for each within the first year of life (85.3%, 1,137 operations out of
year are provided in ►Supplementary Appendix D, available 1,333 patients admitted with the diagnosis); 65.5% of neo-
in the online version). The majority of HD patients were nates with HD had a primary repair.
diagnosed and admitted to the hospital within the first week
of life (n ¼ 970, 72.8%). The majority of patients were males. Main Outcomes in the Specialist Pediatric Centers
The percentage of HD patients who were white was in line The 1-year all-cause mortality rate was 2.8% (n ¼ 37). The
with that expected for England (n ¼ 920, 69.0%). Twenty- overall readmission rate was 70.2% with a significant differ-
seven per cent of patients belonged to the most deprived ence between primary and multiple-staged repair (79.9 vs.
socioeconomic quintile. The most common comorbidity was 90.1%, p < 0.01). The 1-year overall reoperation rate was
major cardiac anomalies (n ¼ 163, 12.2%). 11.9% with no difference between groups.
Table 1 Demographic characteristics of neonates admitted to the specialist pediatric surgical centers in England
All patients admitted All patients operated (N ¼ 1,137) p-Value

(N ¼ 1,333)
Primary repair (N ¼ 874) Multiple staged (N ¼ 263)
Age at admission
<1 d 166 (12.5) 115 (13.2) 29 (11.0) NS
1–6 d 970 (72.8) 630 (72.1) 200 (76.0) NS
7–28 d 197 (14.8) 129 (14.8) 34 (12.9) NS
Sex
Male 991 (74.3) 656 (75.1) 192 (73.0) NS
Female 341 (25.6) 218 (24.9) 70 (26.6) NS
Missing 5 (–) 0 (0.0) 5 (–) NS
Ethnic group
White 920 (69.0) 592 (67.7) 192 (73.0) NS
White mixed 39 (2.9) 28 (3.2) 5 (–) NS
Asian or Asian British 73 (5.5) 43 (4.9) 19 (7.2) NS
Black or black British 55 (4.1) 44 (5.0) 5 (–) NS
Chinese and other 44 (3.3) 28 (3.2) 11 (4.2) NS
Missing 202 (15.2) 139 (15.9) 32 (12.2) NS
Carstairs index
1 192 (14.4) 118 (13.5) 37 (14.1) NS
2 189 (14.2) 120 (13.7) 39 (14.8) NS
3 239 (17.9) 153 (17.5) 46 (17.5) NS
4 270 (20.3) 192 (22.0) 44 (16.7) NS
5 366 (27.5) 238 (27.2) 82 (31.2) NS
Missing 77 (5.8) 53 (6.1) 15 (5.7) NS
Comorbiditiesa
Major cardiac anomalies 163 (12.2) 107 (12.2) 36 (13.7) NS
Chromosomal abnormalities 133 (10.0) 93 (10.6) 22 (8.4) NS
Other low birth weight 34 (2.6) 23 (2.6) 10 (3.8) NS
Other preterm infants 40 (3.0) 73 (8.4) 10 (3.8) NS
1-y outcomes
Mortality 37 (2.8) 23 (2.6) 5 (–) NS
Readmissions 936 (70.2) 699 (79.9) 237 (90.1) < 0.01
Reoperations 159 (11.9) 124 (14.2) 35 (13.3) NS
Abbreviation: NS, not significant.

Note: (–) denotes the percentage has been suppressed due to low numbers.
a
Comorbidities present in <5 cases were not presented in the table as not statistically relevant.

Associations between patient/hospital characteristics, difference in 1-year cumulative reoperation between prima-
30-day emergency readmission, and 1-year cumulative re- ry and multiple-staged repair (14.2 vs. 16.6, p ¼ 0.718). In
operation are shown in ►Table 2. Patients receiving primary addition, patients who were admitted electively had a sig-
repair were less likely to be readmitted as an emergency to a nificantly lower reoperation rate.
hospital within 30 days from discharge (36% compared with There was variation between centers in the main out-
49% for multiple staged, p ¼ 0.004). There was no significant comes. In particular, emergency readmissions ranged from
10 to 48% and reoperation rates from 5 to 28% among
Table 2 Detailed analysis of associations between patient/ different specialist pediatric surgical centers.
hospital characteristics and main outcomes: 30-day emergency
readmission and 1-year cumulative reoperation Secondary Outcomes: Cumulative Length of Stay and
Volume–Outcomes
30-d emergency 1-y cumulative On average, neonates with HD had a cumulative length of
readmission reoperation
stay of 40 days regardless of operation type (primary repair:
Number (%a) Number (%a) mean ¼ 40.9, standard deviation [SD] ¼ 53.8; multiple-
Age at admission p ¼ 0.071 p ¼ 0.598 staged repair: mean ¼ 40.5, SD ¼ 52.5). There was no signifi-
<1 d 42 (29.2) 24 (16.7) cant hospital volume and outcomes relationship found
(►Table 2).
1–6 d 305 (36.7) 112 (13.5)
7–28 d 68 (41.7) 23 (14.1)
Adjusted Logistic Regression
Sex p ¼ 0.75 p ¼ 0.746 ►Table 3 shows the association between procedure type and
Male 310 (36.6) 122 (14.4) outcomes. All the logistic regression models were adjusted
Female 105 (36.5) 37 (12.8) for age, gender, ethnic group, socioeconomic deprivation
Missing 0 (0) 0 (0) (Carstairs index), comorbidity, and year of admission. Proce-
dure type and hospital volume were included as independent
Ethnic group p ¼ 0.118 p ¼ 0.291
variables for their associations with outcomes. Procedure
White 275 (35.1) 115 (14.7) type significantly predicted readmission in HD. There were
White mixed 15 (NA) 5 (NA) no significant associations between socioeconomic depriva-
Asian or Asian British 16 (25.8) 10 (16.1) tion, year of admission, and the presence of comorbidity with
Black or black British 20 (NA) 9 (NA) postoperative readmission and HD. The odds of readmission
within 1 year for neonates who underwent a primary repair
Chinese and other 16 (41.0) 5 (NA)
were 92% lower than the ones who received a multiple staged
Missing 73 (42.7) 21 (12.3) (OR ¼ 0.08, 95% confidence interval: 0.06–0.11, p < 0.001).
Carstairs index p ¼ 0.188 p ¼ 0.493
1 53 (34.2) 25 (16.1)
Discussion
2 65 (40.9) 28 (17.6)
Main Findings
3 82 (41.2) 22 (11.1)
This is the largest study to report 1-year outcomes in HD. The
4 91 (38.6) 34 (14.4)
numbers of HD cases born in England has been stable over
5 102 (31.9) 40 (12.5) the study period with an incidence of 0.91 per 5,000 live
Missing 22 (33.4) 10 (14.7) births. A primary repair was performed on 65.5% of HD
Admission method p ¼ 0.783 p < 0.001 neonates treated in SPSC. After adjusting for comorbidities,
the primary repair was associated with lower odds of read-
Elective 5 (NA) 5 (NA)
mission (8%) and shorter cLOS (38%) compared with multi-
Emergency 192 (37.4) 93 (18.1)
ple-staged repair.
Others 218 (35.9) 65 (10.7) We did not find a significant hospital volume–outcome
Procedure type p ¼ 0.004 p ¼ 0.718 relationship in neonatal colorectal surgery.
Primary repair 312 (35.7) 124 (14.2)
Comparison with Other Literature
Multiple-staged repair 103 (48.8) 35 (16.6)
A primary repair has become the gold standard treatment
Hospital volume p ¼ 0.128 p ¼ 0.944
for HD whenever rectal washouts are effective to decom-
High 199 (32.9) 84 (15.0) press the functionally obstructed bowel at birth. The rate of
Medium 140 (40.5) 48 (13.0) primary repair for HD in SPSC in England was in line with
Low 76 (37.6) 27 (13.4) the data published from recent United Kingdom and Ireland
(64%) and United States (77.2%) cohorts.15,16 This has been a
Note: The p-values are the results of chi-square test assessing the associ-
stable trend in the last two decades, in particular, because
ations. Percentages (a) are the percentage of patients in each group who
received an operation (available in ►Table 1). Where any data has had to of the widespread of minimally invasive surgery and the
be suppressed due to small numbers, the percentage are not available (NA), adoption of novel transanal pull-through technique to
although the chi-square test was performed on the actual data. repair HD.17

Table 3 Adjusted logistic regression analysis for Hirschsprung’s disease to define associations with readmission and cLOS
Readmission More cLOS

Odds ratio 95% CI Pr > chi-square Odds ratio 95% CI p-Value
Age 0.510 0.572
<1 d 1.00 1.00
1–6 d 0.66 0.33–1.34 0.252 1.31 0.79–2.18 0.294
7–28 d 0.66 0.29–1.49 0.317 1.31 0.72–2.37 0.378
Sex 0.997 0.928
Male 1.00 1.00
Female 1.02 0.67–1.54 0.942 1.06 0.78–1.46 0.699
Missing >999.99 0.987 <0.01 0.982
Ethnic group 0.963 0.026
White 1.00 1.00
White mixed 1.30 0.38–4.44 0.677 1.20 0.52–2.79 0.673
Asian or Asian British 1.23 0.54–2.85 0.623 1.56 0.88–2.78 0.129
Black or black British 1.33 0.46–3.84 0.601 0.57 0.25–1.32 0.191
Chinese and other ethnic group 1.27 0.43–3.76 0.662 1.46 0.71–3.01 0.306
Missing 0.92 0.56–1.53 0.759 1.71 1.18–2.47 0.004
Carstairs index 0.779 0.963
1 1.00 1.00
2 1.17 0.57–2.40 0.667 1.06 0.63–1.79 0.832
3 0.92 0.48–1.79 0.811 1.16 0.70–1.91 0.563
4 1.03 0.53–1.98 0.931 1.08 0.67–1.77 0.747
5 0.85 0.46–1.56 0.605 1.03 0.65–1.64 0.905
Missing 0.59 0.22–1.55 0.282 1.37 0.64–2.90 0.416
Year of admission 0.950 0.643
2003 1.00 1.00
2004 1.41 0.51–3.93 0.513 1.45 0.71–2.99 0.312
2005 1.42 0.53–3.77 0.484 0.97 0.47–1.98 0.929
2006 1.05 0.41–2.67 0.922 0.85 0.41–1.75 0.651
2007 0.95 0.36–2.48 0.912 1.02 0.49–2.14 0.959
2008 1.13 0.40–3.17 0.822 1.06 0.49–2.31 0.883
2009 1.10 0.42–2.86 0.853 1.12 0.55–2.31 0.752
2010 1.63 0.59–4.56 0.348 0.78 0.37–1.62 0.499
2011 1.01 0.40–2.52 0.990 0.80 0.39–1.64 0.549
2012 0.81 0.33–2.00 0.652 0.80 0.39–1.65 0.543
2013 0.86 0.34–2.15 0.740 1.01 0.49–2.08 0.970
2014 1.11 0.43–2.90 0.831 0.64 0.30–1.36 0.245
2015 1.60 0.55–4.63 0.390 0.64 0.29–1.40 0.259
Comorbidity 0.345
Yes 1.00 1.00
No 1.17 0.85–1.62 0.345 0.89 0.68–1.17 0.389
Procedure type <0.001 <0.001
Multiple stage 1.00 1.00
Single stage 0.08 0.06–0.11 <0.001 0.38 0.28–0.52 <0.001
Hospital volume 0.844 0.170
Low 1.00 1.00
Medium 0.87 0.50–1.50 0.606 0.75 0.50–1.12 0.156
High 0.96 0.58–1.60 0.872 0.70 0.49–1.02 0.062
Abbreviations: CI, confidence interval; cLOS, cumulative length of hospital stay.

Sulkowski et al reported HD outcomes from a large number of cases per year (5.6 for HD) and the average
American administrative database.15 They showed that number of consultant surgeons in each center is 6 to 8.
patients undergoing multiple-staged repair had significantly This has significant implication for practice, as the volume of
higher rates of readmissions (58.5 vs. 37.9%) and additional neonatal colorectal cases is probably not enough to maintain
operations (38.7 vs. 26%) compared with primary repair. We surgical skills in every hospital. Some specialist pediatric
did find similar results for readmission but not for reopera- centers have already developed subspecialization having
tion, where the primary repair had no significant difference shown higher number of cases per surgeon, multidisciplin-
compared with multiple-stage repair (14.2% in the primary ary clinics, a better structured postoperative follow-up, and
vs. 16.6% in the multiple-stage repair, p ¼ 0.718). The same transition of care. More qualitative research data are also
study also reported that patients in the multiple-staged essential to evaluate the need for subspecialization or re-
repair had higher rates of SSI, postoperative enterocolitis, gionalization of care.
small bowel obstruction, and higher total adjusted hospital
costs for overall inpatient care. Several studies have con- Strengths and Limitations
firmed safety and cost-effectiveness of primary repair in HD This is one of the first studies to summarize the incidence of
to reinforce the indication for the use of this technique when key neonatal malformations in England. We analyzed a large
clinically indicated.2,18,19 database, which includes all infants born in NHS hospitals in
The complication rate associated with a neonatal colosto- England. We have considered key 1-year outcomes for infants
my is high, and it has been reported between 23 and 30%.20,21 born with the main colorectal congenital anomalies and
The main complications are wound dehiscence, prolapse, provide important benchmark data for clinicians and parents.
intestinal obstruction, skin rash, and bleeding. In developing There were several limitations related to the nature of
countries, colostomy formation for anorectal malformations administrative data. First, the types of analyses are limited by
was associated with significantly high mortality (11.5%) and the ICD-10 codes that can capture only some of the clinically
surgical site infection (19.7%) rates.22 In addition, we should important variables and outcomes. Second, the data may
not underestimate the psychosocial impact and the manage- have coding errors for several reasons. Third, it was not
ment of an infant with a colostomy for the family. possible to differentiate for severity and types of congenital
Analysis of costs for primary versus multiple-stage repair defects. Patients with similar levels of disease severity may
in HD has shown similar or reduced costs in favor of the receive the same type of procedures, which could have
primary repair.18,19 In particular, reduced length of stay and contributed to the differences in outcomes that we observed.
readmission can drive cost reduction and better patient and
parent experience. Further research on the cost-effectiveness
Conclusion
of primary repair should be performed to better define
possible savings and impact on the quality of care. Primary repair for HD had better postoperative outcomes
than multiple staged. Neonatal colorectal surgeons should
The Implication for Clinical Practice consider the use of a primary repair whenever clinically
Whenever possible, a primary repair for HD should be possible and safe for the patient.
performed to achieve a shorter LOS, lower readmission
rate, no colostomy complications, better cosmesis, and over- Conflict of Interest
all better patient/parent experience. The findings from this None declared.
article have significant implication for surgical practice as
there are many surgeons still using the multiple-stage repair Acknowledgments
as their preferred treatment for the repair of these congenital The Dr Foster Unit at Imperial College London is affiliated
anomalies. For HD, the primary repair has become the with the National Institute of Health Research (NIHR)
technique of choice over the last decade, but there is still a Imperial Patient Safety Translational Research Centre.
significant variation in the use of multiple-staged repair.23 A The NIHR Imperial Patient Safety Translational Centre is
published survey from the United States showed that most of a partnership between the Imperial College Healthcare
the surgeons (85%) would advocate a primary repair without NHS Trust and Imperial College London. The Dr Foster Unit
colostomy.24 We found that in England, one-third of HD at Imperial College London is grateful for support from the
infants had a colostomy formed as part of their treatment, NIHR Biomedical Research Centre funding scheme. The
and this number can be improved. views and opinions expressed herein are those of the
Based on our results and the evidence presented from the authors and do not necessarily reflect those of the NHS,
literature, an experienced pediatric colorectal surgeon NIHR, MRC, CCF, NETSCC, HSR program, or Department of
should adopt a primary repair in most of the HD cases as Health.
the incidence of long segment is only 5 to 10%.25
There is still an ongoing debate about the benefits of
References
subspecialization of care for pediatric colorectal surgery.
1 de Lorijn F, Boeckxstaens GE, Benninga MA. Symptomatology,
We could not find a positive relationship between hospital pathophysiology, diagnostic work-up, and treatment of Hirsch-
volume and outcomes in HD. We need to highlight the fact sprung disease in infancy and childhood. Curr Gastroenterol Rep
that even high-volume hospitals in England had a small 2007;9(03):245–253

2 Langer JC, Durrant AC, de la Torre L, et al. One-stage transanal 14 Wang Y, Honeyford K, Aylin P, Bottle A, Giuliani S. One-year
Soave pullthrough for Hirschsprung disease: a multicenter expe- outcomes for congenital diaphragmatic hernia. BJS Open 2019;3
rience with 141 children. Ann Surg 2003;238(04):569–583 (03):305–313
3 Bradnock TJ, Knight M, Kenny S, Nair M, Walker GM; British 15 Sulkowski JP, Cooper JN, Congeni A, et al. Single-stage versus multi-
Association of Paediatric Surgeons Congenital Anomalies Surveil- stage pull-through for Hirschsprung’s disease: practice trends and
lance System. Hirschsprung’s disease in the UK and Ireland: outcomes in infants. J Pediatr Surg 2014;49(11):1619–1625
incidence and anomalies. Arch Dis Child 2017;102(08):722–727 16 Bradnock TJ, Knight M, Kenny S, Nair M, Walker GM; British
4 Haricharan RN, Georgeson KE. Hirschsprung disease. Semin Association of Paediatric Surgeons Congenital Anomalies Surveil-
Pediatr Surg 2008;17(04):266–275 lance System (BAPS-CASS). The use of stomas in the early manage-
5 De La Torre L, Langer JC. Transanal endorectal pull-through for ment of Hirschsprung disease: findings of a national, prospective
Hirschsprung disease: technique, controversies, pearls, pitfalls, cohort study. J Pediatr Surg 2017;52(09):1451–1457
and an organized approach to the management of postopera- 17 Swenson O. Hirschsprung’s disease: a review. Pediatrics 2002;109
tive obstructive symptoms. Semin Pediatr Surg 2010;19(02): (05):914–918
96–106 18 Hackam DJ, Superina RA, Pearl RH. Single-stage repair of Hirsch-
6 Zani A, Eaton S, Morini F, et al; EUPSA Network Office. European sprung’s disease: a comparison of 109 patients over 5 years. J Pediatr
Paediatric Surgeons’ Association survey on the management of Surg 1997;32(07):1028–1031
Hirschsprung disease. Eur J Pediatr Surg 2017;27(01):96–101 19 Shinall MC Jr, Koehler E, Shyr Y, Lovvorn HN III. Comparing cost
7 Bandi AS, Bradshaw CJ, Giuliani S. Advances in minimally invasive and complications of primary and staged surgical repair of
neonatal colorectal surgery. World J Gastrointest Surg 2016;8 neonatally diagnosed Hirschsprung’s disease. J Pediatr Surg
(10):670–678 2008;43(12):2220–2225
8 Strebeck J. Acute & chronic wounds: current management con- 20 van den Hondel D, Sloots C, Meeussen C, Wijnen R. To split or not
cepts. AORN J 2007;86(04):677–678 to split: colostomy complications for anorectal malformations or
9 Faiz O, Brown T, Bottle A, Burns EM, Darzi AW, Aylin P. Impact of Hirschsprung disease: a single center experience and a systematic
hospital institutional volume on postoperative mortality after major review of the literature. Eur J Pediatr Surg 2014;24(01):61–69
emergency colorectal surgery in English National Health Service 21 Liechty ST, Barnhart DC, Huber JT, Zobell S, Rollins MD. The
Trusts, 2001 to 2005. Dis Colon Rectum 2010;53(04):393–401 morbidity of a divided stoma compared to a loop colostomy in
10 NHS Digital. Available at: http://content.digital.nhs.uk/hes. patients with anorectal malformation. J Pediatr Surg 2016;51
Accessed January 31, 2018 (01):107–110
11 Burns EM, Bottle A, Aylin P, Darzi A, Nicholls RJ, Faiz O. Variation in 22 Chirdan LB, Uba FA, Ameh EA, Mshelbwala PM. Colostomy for high
reoperation after colorectal surgery in England as an indicator of anorectal malformation: an evaluation of morbidity and mortali-
surgical performance: retrospective analysis of Hospital Episode ty in a developing country. Pediatr Surg Int 2008;24(04):407–410
Statistics. BMJ 2011;343:d4836 23 Langer JC, Fitzgerald PG, Winthrop AL, et al. One-stage versus
12 Best KE, Glinianaia SV, Bythell M, Rankin J. Hirschsprung’s disease two-stage Soave pull-through for Hirschsprung’s disease in the
in the North of England: prevalence, associated anomalies, and first year of life. J Pediatr Surg 1996;31(01):33–36
survival. Birth Defects Res A Clin Mol Teratol 2012;94(06): 24 Keckler SJ, Yang JC, Fraser JD, et al. Contemporary practice
477–480 patterns in the surgical management of Hirschsprung’s disease.
13 Burns EM, Bottle A, Almoudaris AM, et al. Hierarchical multilevel J Pediatr Surg 2009;44(06):1257–1260
analysis of increased caseload volume and postoperative outcome 25 Grosfeld J, O’Neill J, Coran A, Fonkalsrud E. Pediatric Surgery, 6th
after elective colorectal surgery. Br J Surg 2013;100(11):1531–1538 ed. Mosby; 2006:1514–1559

Original Article
Indications, Surgical Complications, and Long-Term

Outcomes in Pediatric Esophageal Reconstructions
with Pedicled Jejunal Interposition Graft
Antti Koivusalo1 Janne Suominen1 Jukka Salminen2 Mikko Pakarinen1
1 Department of Pediatric Surgery, Children’s Hospital, University of Address for correspondence Antti Koivusalo, MD, PhD, Department
Helsinki, Helsinki, Finland of Pediatric Surgery, Children’s Hospital, University of Helsinki,
2 Department of Pediatric Cardiac Surgery, Children’s Hospital, Stenbäckinkatu 11 PL 281, Helsinki 00290, Finland
University of Helsinki, Helsinki, Finland (e-mail: antti.koivusalo@hus.fi).
Eur J Pediatr Surg
Abstract Introduction Several surgical techniques are available for pediatric esophageal
reconstruction. We started to use pedicled jejunum interposition graft (PJIG) because
other techniques had significant long-term complications. In this retrospective study,
the indications, surgical complications, and long-term outcomes were assessed in
patients with PJIG.
Materials and Methods With ethical consent, we reviewed the hospital records of 14
patients (7 females) who from 2005 to 2019 underwent a total of 16 esophageal
reconstructions with PJIG.
Results Median age at PJIG was 1.6 (range: 0.2–15) years. Underlying conditions were
esophageal atresia (EA) (n ¼ 11) or native esophagus lost by trauma or infection
(n ¼ 3). Eight patients with EA underwent PJIG as primary reconstruction and three as a
rescue operation after complications in primary repair. Significant surgical complica-
tions occurred in 43% of patients. Major reoperations in six (43%) patients included
resection and reanastomosis of strictured proximal PJIG (n ¼ 1) and redo PJIG after
failure of the first operation (n ¼ 2). Surgical mortality was nil. After a median follow-up
of 6.5 (range: 0.7–14) years, 13 (93%) patients survived, and 1 died of congenital heart
disease. PJIG failed in three (23%) survivors of whom two underwent graft removal
Keywords because of life-threatening aspiration and one did not start oral feeds at all. Ten
► esophageal survivors (77%) have full enteral feeds. Respiratory function in the survivors is
reconstruction satisfactory. Two patients have moderate and three mild gastroesophageal reflux
► jejunal interposition symptoms.
► surgical Conclusion PJIG was a functional option for a variety of conditions that required
complications esophageal reconstruction. However, significant early and late complications required
► long-term outcomes major surgical revisions.
Introduction gap EA, successful repair with the native esophagus is possible
at least in half of the patients1 and ingenious techniques such
In children, esophageal atresia(EA) or loss of esophagus be- as esophageal lengthening by traction may enable the use of
cause of trauma such as ingestion of caustic fluids is the most native esophagus in still greater part of patients with long-gap
common indications for esophageal reconstruction. In long- EA.2,3 Sometimes, however, the shortness or poor quality of

December 3, 2019
Pediatric Esophageal Reconstructions with Pedicled Jejunal Interposition Graft Koivusalo et al.
the available esophageal tissue precludes repair with native accordingly at a slower pace. Antithrombotic medication
esophagus and esophageal reconstruction is required. In our was not routinely given.
center, peedicled jejunum interposition graft (PJIG) was
adopted as the first option for esophageal reconstruction. Diagnosis and Management of Surgical Complications
We chose PJIG at 2005 mainly because of unsatisfactory Anastomotic leakage was diagnosed with the help of contrast
long-term results after Gavriliu–Heimlich type reversed gas- radiographs or with contrast computed tomography (CT)
tric tube4 reconstruction and because contemporary pub- scan. Gross leakage was managed surgically, whereas con-
lished results of JI were promising.5 In this article, we servative treatment with drains and antibiotics was
describe our experience with PJIG. attempted in a stable patient with moderate or minor
leakage. Anastomotic strictures were diagnosed with endos-
copy with the help of oral contrast radiographs. The mainstay
in management of strictures was endoscopic hydrostatic
Institutional ethical consent (447/E7/2004) was obtained. dilatation. We graded the severity GER by clinical symptoms
We reviewed the hospital records pediatric patients who such as regurgitation, vomiting, and aspiration. Assessment
underwent esophageal reconstruction from 2005 to 2019. with pH/impedance monitoring was not used. Assessment of
We recorded the indications for reconstruction: early surgi- GER was completed with endoscopy and contrast radiograph
cal complications such as leakage or stricture. Late compli- when clinically applicable. All patients were included in
cations included gastroesophageal reflux (GER), recurrent endoscopic follow-up program that included at least 1 yearly
stricture, pulmonary complications, and loss of the graft. The endoscopy for the first 3 years and after that endoscopic
method of enteral feeding was classified as full oral or follow-up at every 5 years or whenever the need arises.
feeding dependent on gastrostomy or jejunostomy with or Endoscopy included esophageal mucosal biopsies to detect
without oral feeding. esophagitis and dysplasia, which were graded according to
In patients with EA, reconstruction as a primary means of Hetzel et al7 and Ismail-Beigi et al.8 In addition, all patients
repair was considered in patients with long-gap EA (Gross were under surveillance by a pediatric pulmonologist, a
types A and B). Gap length was measured in fluoroscopy- dietician, and a speech therapist.
guided endoscopy from above into the proximal pouch and Data were expressed in median and range.
through gastrostomy into distal pouch, details of the method
have been described earlier.1 Reconstruction was considered
Results
if the gap length was four or more vertebral bodies at age
from 6 weeks. The same method of gap measurement was In total, 14 patients (7 females) were included in the study.
also used in patients with loss of esophagus from another Median age at reconstruction was 1.6 (range: 0.2–15) years.
cause than EA whenever applicable. Reconstruction was a Clinical data are outlined in ►Table 1. Median duration of
preferred option in patients with cervical esophagostomy or hospitalization was 30 (range: 10–187) days. Median age in
with an earlier unsuccessful attempt on end-to-end repair. patients who underwent reconstruction as the primary
procedure for EA (n ¼ 8) was 0.6 (range: 0.2–2.0) years, in
Pedicled Jejunum Interposition Graft patients who underwent reconstruction as a rescue proce-
The technique of PJIG has been published earlier by Katsura dure after failed primary repair of EA (n ¼ 3) was 3.4 (range:
et al (1958)6 and Bax and van der Zee (2007).5 We did not 1.8–4.9) years, and in patients with other indications (n ¼ 3)
perform pyloroplasty or pyloromyotomy if no signs of ob- was 10 (range: 3.4–15) years. Two of 14 patients with PJIG
struction occurred during preoperative gastrostomy feeding. underwent redo PJIG after failure of the first reconstruction,
Alternative technique included placement PJIG in retroster- thus the total number of PJIG reconstructions was 16. The
nal space through a midline sternotomy. If retrosternal route location of the PJIG after final reconstruction was orthotopic
is used, midline division of neck muscles and minor osteot- or right posterior mediastinum in 11 patients and retroster-
omy of the upper part of the sternum may be required to nal space in 3. During the same study period from 2005 to
facilitate unhindered passage of the upper esophagus in the 2019, we performed four esophageal reconstructions to four
midline. At the caudal end of the sternotomy, a limited patients with other techniques than PJIG.
division of the anterior margin of the diaphragm may be
needed to prevent kinking of the caudal part of the graft. To Esophageal Atresia, Primary Reconstruction
prevent gliding, the graft may be fixed into the anterior Eight patients underwent PJIG as the primary procedure for
margin of the diaphragm. EA. Gross types of esophageal atresia were A (n ¼ 5, patients
We performed a routine contrast roentgenogram from 5 1–5), B (n ¼ 2, patients 6–7), and D (n ¼ 1, patient 8). Median
to 7 days after the reconstruction, and if no signs of compli- gap length was 6.5 (range: 4.0–8.0) thoracic vertebrae. In all
cations were detected, we started feeding first by gastro- eight patients, jejunal interposition graft was placed in
stomy and then orally. The vitality of the graft and orthotopic position. Early surgical complications included
anastomotic patency was checked with endoscopy 3 to distal anastomotic leakage in one patient (no. 3) and perfo-
4 weeks postoperatively. In case of frequent or copious ration of the jejunal graft by a misplaced nasogastric tube in
postoperative vomiting, we performed an omega feeding another patient (no. 4). The anastomotic leakage (no. 3)
jejunostomy and increased oral and gastrostomy feeding healed with external drain age and antibiotics, and the

Table 1 Clinical data of the 14 patients who underwent esophageal reconstruction with PJIG
Patient Sex Major Age at Indication PREOP Location Distal Complications Major anastomotic
associated PJIG (y) CO of PJIG anastomosis reoperations
anomaly
1 F None 2.0 LGOA No RPM Distal esophagus SPA Resection and
reanastomosis
2 F 21-trisomy 0.9 LGOA No RPM Distal esophagus SPA Stenting
3 M 21-trisomy, 0.2 LGOA No RPM Distal esophagus Leakage, SDA Stenting
VSD PDA
4 F 21-trisomy, 1.0 LGOA Yes RPM Distal esophagus Leakage Suture closure
PH
5 F None 0.4 LGOA No RPM Gastric cardia None None
6 M Muscular 0.2 LGOA No RPM Distal esophagus None None
hypotonia
7 M Tetralogy 0.2 LGOA No RPM Distal esophagus None None
of Fallot
8 M Laryngeal 2.1 LGOA Yes RPM Distal esophagus None None
atresia
9 F None 1.2 SROA No RPM Distal esophagus Leakage None
10 M None 5.5 SROA No RPM Gastric cardia None None
11 M None 2.6 SROA Yes (RPM) Distal esophagus Leakage Graft removed
3.4 RE PJIG Yes RS Gastric fundus None None
12 M None 2.4 LOCI Yes RPM Distal esophagus None None
13 F None 15.6 LOCI Yes RS Gastric fundus None None
14 F None 9.6 LOCI Yes RPM Distal esophagus Graft necrosis, Graft removed
leakage
10.4 RE PJIG Yes RS Gastric fundus None None
7/14 (50%) 6/14 (43%)
Abbreviations: LGOA, long-gap esophageal atresia; LOCI, loss of esophagus, corrosion injury or infection; PDA, patent ductus arteriosus; PH,
pulmonary hypertension; PJIG, pedicled jejunal interposition graft; PREOP CO, preoperative cervical esophagostomy; REPJIG, redo PJIG; RPM, right
posterior mediastinum; RS, retrosternal; SDA, stricture of distal anastomosis; SPA, stricture of proximal anastomosis; SROA, secondary
reconstruction, esophageal atresia; VSD, ventricular septal defect.
perforated graft (patient 4) was successfully suture closed via underwent also laparotomy for adhesion occlusion, reloca-
rethoracotomy. Three patients (nos. 1, 2, and 6) developed a tion of the feeding jejunostomy because of abdominal wall
recurrent stricture of the upper esophagus–graft anastomo- site infection and had long periods of parenteral nutrition
sis and one patient (no. 3), a stricture of the distal graft– with associated cholestasis. In patient 3, there was leakage
esophagus anastomosis. In patient 1, the stricture was and subsequent stricture of the distal graft–esophagus anas-
recalcitrant to repeated endoscopic dilatations and resection tomosis. The leakage was managed by external drainage and
of the stricture and reanastomosis was performed; after still the stricture with endoscopic dilatations and placement of
more anastomotic dilatations, the new anastomosis eventu- ELLA stent, but the patient died of cardiac disease. Patient 5
ally remained patent and full oral feeding was achieved. had no distal esophageal stump at all and we performed
Patient 2 developed also proximal stricture that was recalci- successful PJIG with distal anastomosis between graft and
trant to dilatations, and an attempt of stenting with ELLA gastric cardia. After a period with feeding jejunostomy, she
stent (ELLA-CS, Hradec Králové, Czech Republic) was com- started full oral intake in 5 months. Patients 7 and 8 under-
plicated by stent usurpation into retroesophageal aberrant went successful PJIG and were on full oral feeds 14 and
right subclavian artery which resulted in major hemorrhage 24 months after PJIG. Of eight patients with primary recon-
and vascular prosthetic surgery. Full oral feeds were never struction for long-gap EA, a total of five (58%) had significant
achieved and because of copious vomiting, she eventually postoperative complications.
had feeding jejunostomy. In patient 6, the proximal stricture
was managed with repeated endoscopic dilatations. His Esophageal Atresia, Secondary Reconstruction
recovery was complicated by copious vomiting and aspira- PJIG was used as secondary reconstruction in two patients
tion pneumoniae with abscess of the left lung. He did not with type C (patients 9 and 10) and one with type F (patient
tolerate gastrostomy feeding and even with jejunostomy 11) congenital stenosis) EA. After primary repair, patient 9
feeding, the vomiting continued though less frequently. He developed a recalcitrant anastomotic stricture. After over 30

endoscopic dilatations and fundoplication, the patient even- although preserving good circulation, was too short for
tually underwent resection of the 4 cm long stricture. The reanastomosis, and consequently, the graft was excised
resulting esophageal defect was not amenable to new end- and the cervical esophagostomy was redone. In the second
to-end anastomosis, and therefore, PJIG was used. A minor attempt of reconstruction, we managed to extract another
postoperative leakage was managed with external drainage jejunal graft with adequate vascular arcades. The graft was
and thereafter the esophageal conduit remained patent and placed in to the retrosternal position via midline sternotomy
the patient eventually reached full oral feeds. Patient 10 also and anastomosed with the upper esophagus and the fundus
developed anastomotic stricture after the initial primary of the stomach (►Tables 1 and 2). Postoperative recovery was
anastomosis. He underwent fundoplication and resection uneventful and the patient regained full feeds in 3 months.
of the anastomotic stricture twice, but still required repeated
endoscopic dilatations. Although a seemingly adequate dila- Other Indications
tation of 15 to 16 mm was achieved, dysphagia recurred after Three reconstructions with PJIG were performed after loss of
every 2 to 3 weeks. After well over 120 endoscopic dilatations esophagus after perforation and mediastinitis after battery
by the age of 5 years, the patient finally underwent resection ingestion (patient 12), ingestion of caustic fluid (patient 13),
of the lower esophagus from above the level of the former and infected bronchogenic cyst (patient 14). Initially, all
esophageal anastomosis down to gastric cardia. Esophageal three patients underwent emergency resection of esophagus
conduit was restored with PJIG through right posterior and cervical esophagostomy. Patient 12 underwent recon-
mediastinum with the lower anastomosis between the graft struction with PJIG anastomosed with the upper esophagus
and the gastric cardia. The distal esophagus was abandoned and the remaining stump of lower esophagus. Recovery was
because of the risk of poor motility. Postoperative course was uneventful and the patient reached full oral feeds in
uneventful and he rapidly achieved full oral feeds and no 3 months. In patient 13, the lower esophagus and the gastric
more dilatations were needed. In patient 11, congenital cardia were resected and reconstruction was made with PJIG
esophageal stenosis was initially managed with serial endo- placed retrosternally via midline sternotomy and anasto-
scopic dilatations. After the third session, the patient had mosed with upper esophagus and gastric fundus. Recovery
esophageal perforation that was initially managed with chest was uneventful and return to full oral feeds occurred within
drainage. Eventually, mediastinitis developed and the pa- 1 month. In patient 14, the first reconstruction was compli-
tient underwent emergency resection of the distal esopha- cated by graft necrosis possible because of recurred infection
gus and cervical esophagostomy. First attempt of esophageal of a remaining bronchogenic cyst. Consequently, the graft
reconstruction with PJIG was complicated by gross leakage. was excised and the cervical esophagostomy was redone. In
In reoperation, it was found that after excision of the ische- the second reconstruction, it was possible to devise a new
mic distal part of the upper esophagus, the jejunal graft, jejunal graft with adequate vascular arcade, and the PJIG was
Table 2 Long-term outcome of the 14 patients who underwent esophageal reconstruction with PJIG
Patient Follow-up (y) FJ Outcome

1 12.4 No Full oral feeds
2 5.2 Yes No oral feeds
Graft removed because of constant aspiration at 4.4 y
3 1,1 No Oral feeds started, died of heart failure at the age of 1.4 y
4 7.9 Yes No oral feeds
Severe regurgitation, retardation, and pulmonary hypertension
5 1.6 Temporary Full oral feeds
6 9.8 Yes Reached full oral feeds,
Graft removed because of constant aspiration at 8.9 years
6.5 (range: 0.7–14.0) Survived 13/14 (93%)
Full oral feeds 10/13 survivors (77%)
Abbreviations: FJ, feeding jejunostomy; PJIG, pedicled jejunal interposition graft.

placed retrosternally via midline sternotomy and anasto- following reasons: (1) A patient with giant omphalocele to
mosed with upper esophagus and gastric fundus. Recovery whom we had to perform esophageal reconstruction for
was uneventful and the patient was with full oral feeds long-gap type A EA and final closure of the abdominal wall
within 2 months. during the same procedure. We thought that blood circula-
tion in PJIG was too vulnerable to be exposed to the possible
Long-Term Outcomes risk of increased intra-abdominal pressure after closure of
Long-term outcomes are outlined in ►Table 2. In three (23%) the abdominal muscles. Consequently, we resorted to gastric
of 13 survivors, the PJIG eventually failed. Patient 3 died of pull-up with more robust circulation and cervical anasto-
heart failure 1 year after PJIG. Two patients (nos. 2 and 6) mosis only. (2) In a patient with type C EA with failed primary
with PJIG as primary reconstruction for long-gap EA suffered anastomosis. The patient had undergone surgery for small
from constant episodes of aspiration even with jejunal bowel volvulus with strangulation and the mesentery of the
feeding. At the age of 3 years, patient 2 spent more than jejunum was extensively scarred and constricted and after
6 months in intensive care because of recurrent aspiration open cardiac surgery, the retrosternal graft placement was
pneumonia. As her respiratory function was deteriorating not possible. We deemed PJIG too risky and used gastric pull-
rapidly, we decided to remove the jejunal graft and leave the up instead. (3) A patient with type C EA with failed primary
patient with cervical esophagostomy and feeding jejunos- anastomosis, missing right lung, very severe tracheobron-
tomy. Patient 6 achieved full oral feeds and the feeding chomalacia, and a very short high cervical esophagostomy.
ostomies were closed at the age of 7 years, but he continued (4) We used free jejunal graft in a patient with traumatic
to have recurrent aspiration pneumonia and developed tracheoesophageal fistula in whom the esophageal defect
bronchiolitis. At the age of 8 years, his respiratory function was too proximal for management with PJIG.
deteriorated rapidly to less than 25% of age-adjusted referral
values and he needed extra oxygen. We removed his jejunal
Discussion
graft and left him for cervical esophagostomy and feeding
jejunostomy. At surgery, we found a deep sinus or graft– The present series of 16 pedicled jejunal interposition grafts
pulmonary fistula near the proximal anastomosis and the represents 80% of operations for esophageal reconstruction
mucosa of the graft severely atrophied. After removal of the in our center between 2005 and 2019. Initially, we had high
dysfunctional jejunal grafts, both patients 2 and 6 had expectations on the functional advantages of pedicled jejunal
significant improvement of respiratory function and im- interposition graft. The supposed advantages of PJIG includ-
proved weight gain. Patient 2 never achieved full oral feeds ed anastomosis without tension, good circulation, good
and planning of a new esophageal reconstruction remains peristalsis of the graft, less tendency to graft dilatation,
questionable. In patient 6, gastric pull-up may be feasible in and redundancy as the patient grows compared with colon
the future. Patient 4 has significant mental retardation and interposition graft, reflux barrier created by the graft and
has no oral intake. Considering her poor cardiac and pulmo- possibly better resistance to carcinogenic effect of gastric
nary function, we are content to keep her on jejunostomy acid and bile compared with colon interposition graft. In
feeding to avoid aspiration. reality, however, we had some very hard lessons with major
Patients 1, 5, and 7 to 14 are on full oral feeds with surgical complications. The main indication for PJIG was EA
satisfactory or good nutritional status. Respiratory function and many of these patients had significant associated anom-
parameters are satisfactory with forced vital capacity and alies that increased the complexity and risks of the recon-
forced expiratory volume in 1 second values 70 to 90% of the struction. In a country with 5.5 million population with
age-adjusted reference values. annual case load of one to two pediatric esophageal recon-
All survivors with remaining graft (n ¼ 11) have under- structions even with centralization of cases, a steep learning
gone endoscopic surveillance. Conduit patency has remained curve for PJIG is hard to come by.
good in all patients. Biopsies disclosed no pathology in the Compared with the contemporary literature,5,11,12 our
jejunal grafts, and two patients had mild esophagitis in the complication rate (43%) and failure rate (23%) were relatively
proximal native esophagus. Of patients remaining with a high, whereas surgical mortality was nil. Perhaps not unsur-
functioning PJIG, five (45%) have distal graft–stomach anas- prisingly, we gained important experience concerning the
tomosis and thus no antireflux barrier offered by the distal management surgical complications and patient selection.
esophagus. Two of these five patients have GER symptoms We discovered that after an unsuccessful PJIG, it may be
that require elevated position of the upper body while lying possible to develop a second PJIG if the anatomy of the
or sleeping. Three other patients have occasional symptoms proximal vascular arcades of the jejunum is favorable. Place-
of GER and use proton pump inhibitors. ment of the graft in retrosternal position gives the surgeon
excellent field of work for a redo operation. We found also
Patients Who Were Considered Unsuitable for PJIG that in case of a stricture of the proximal esophagus–graft
During the study period 2005 to 2019, we performed esoph- anastomosis, it was possible to resect and perform a new
ageal reconstruction to a total of 18 patients of whom 14 anastomose without undue tension because of the extra
(78%) underwent PJIG and 4 (22%) another type of recon- length and tortuosity of PJIG. Our experience in stenting of
struction including gastric pull-up (n ¼ 3)9 and free jejunal anastomotic strictures was limited. In one patient, sudden
graft (n ¼ 1).10 We preferred gastric pull-up to PJIG for the bleeding after stent usurpation into an aberrant

retroesophageal subclavian artery underlined the need of Conclusion

preoperative assessment of upper body vasculature with CT
scan. Resection and reanastomosis was an alternative to Fourteen pediatric patients underwent esophageal recon-
repeated dilatations and stenting. In patients with copious struction by pedicled jejunal interposition graft. The most
postoperative vomiting, resorting to jejunostomy feeding common indication for reconstruction was EA. Significant
without too much delay seemed advisable. surgical complications were frequent, but surgical mortality
Our series included three patients with undeniable treat- was nil. Lifelong multidisciplinary surveillance for graft
ment failures. These included two patients (nos. 2 and 6) patency and function, respiratory function, and nutritional
with dysfunctioning reconstructions and with constant status is essential for the well-being of the patients.
problems with aspiration. In retrospect, it is obvious that
we treated their problems too long with inadequate meas- Conflict of Interest
ures such as repeated endoscopic dilatations, jejunostomy None declared.
feeding, speech and swallowing therapy, and simply by
waiting for the patients to outgrow their problems. In both
patients, the result was a life-threatening respiratory col- References
lapse. By timely removal of the PJIG, cervical esophagostomy 1 Koivusalo AI, Sistonen SJ, Lindahl HG, Rintala RJ, Pakarinen MP.
and feeding jejunostomy fast recovery of respiratory func- Long-term outcomes of oesophageal atresia without or with proxi-
mal tracheooesophageal fistula - Gross types A and B. J Pediatr Surg
tion followed. The third patient (no. 4) lives in an institution
2017;52(10):1571–1575
for mentally retarded children and from the beginning was 2 Bairdain S, Hamilton TE, Smithers CJ, et al. Foker process for the
not a good candidate for reconstruction. For her, a cervical correction of long gap esophageal atresia: primary treatment
esophagostomy and feeding ostomy could have been at least versus secondary treatment after prior esophageal surgery. J Pediatr
as good choice as esophageal reconstruction. Surg 2015;50(06):933–937
3 van der Zee DC, Gallo G, Tytgat SH. Thoracoscopic traction
Based on the available scientific data, it is difficult to
technique in long gap esophageal atresia: entering a new era.
determine the potential advantages of one reconstruction Surg Endosc 2015;29(11):3324–3330
technique over others.11,12 Although a surgeon may be profi- 4 Heimlich HJ. Reversed gastric tube (RGT) esophagoplasty for
cient and favor one type of esophageal reconstruction,8 it is our failure of colon, jejunum and prosthetic interpositions. Ann
opinion that it is appropriate to choose reconstruction tech- Surg 1975;182(02):154–160
nique according with the special requirements of each indi- 5 Bax NM, van der Zee DC. Jejunal pedicle grafts for reconstruction
of the esophagus in children. J Pediatr Surg 2007;42(02):363–369
vidual patient. To obtain enough jejunal length to reach upper
6 Katsura S, Ishikawa Y, Okayama G. Transplantation of the partially
esophagus especially in the neck area may not be possible with resected middle esophagus with a jejunal graft. Ann Surg 1958;
PJIG6 and gastric pull-up, free jejunal graft, or a pedicled ileal 147(02):146–156
graft13 may offer a better alternative. The choice of reconstruc- 7 Hetzel DJ, Dent J, Reed WD, et al. Healing and relapse of severe
tion with intestinal grafts may also be limited by loss of peptic esophagitis after treatment with omeprazole. Gastroen-
intestine for various reasons. PJIG is a relatively fragile segment terology 1988;95(04):903–912
8 Ismail-Beigi F, Horton PF, Pope CE II. Histological consequences of
of intestine that requires careful handling with good exposure
gastroesophageal reflux in man. Gastroenterology 1970;58(02):
through laparotomy and thoracotomy, whereas gastric pull-up 163–174
can be routed through a manually created posterior mediasti- 9 Kunisaki SM, Coran AG. Esophageal replacement. Semin Pediatr
nal passage and the operation itself may be assisted by Surg 2017;26(02):105–115
laparoscopy and thoracoscopy. In retrospect, we may speculate 10 Nakamura T, Inokuchi K, Sugimachi K. Use of revascularized
jejunum as a free graft for cervical esophagus. Jpn J Surg 1975;
that the problems with recurrent stricture of the proximal PJIG
5(02):92–102
anastomosis could have been avoided by using gastric pull-up
11 Gallo G, Zwaveling S, Groen H, Van der Zee D, Hulscher J. Long-gap
with more robust blood circulation instead. esophageal atresia: a meta-analysis of jejunal interposition, colon
Careful multidisciplinary surveillance of the patients is interposition, and gastric pull-up. Eur J Pediatr Surg 2012;22(06):
mandatory and should extend into adulthood. In addition to 420–425
functional problems with the esophageal reconstruction, re- 12 Gallo G, Zwaveling S, Van der Zee DC, Bax KN, de Langen ZJ,
Hulscher JB. A two-center comparative study of gastric pull-up
spiratory function, growth, and psychosocial aspects need
and jejunal interposition for long gap esophageal atresia. J Pediatr
attention. Although the patients seem to do well enough, the Surg 2015;50(04):535–539
limited respiratory reserve may cause rapid deterioration of 13 Bax NM, Van Renterghem KM. Ileal pedicle grafting for esoph-
lung function. Gallo et al (2017)14 found impaired respiratory ageal replacement in children. Pediatr Surg Int 2005;21(05):
function common among patients with esophageal recon- 369–372
struction of whom patients with gastric pull-up fared worse 14 Gallo G, Vrijlandt EJLE, Arets HGM, et al. Respiratory function after
esophageal replacement in children. J Pediatr Surg 2017;52(11):
than those with PJIG. We agree with the conclusion by Awad
1736–1741
and Jaffray (2017)15 that despite seemingly favorable results, it 15 Awad K, Jaffray B. Oesophageal replacement with stomach: a
is clear that patients with esophageal reconstruction by any personal series and review of published experience. J Paediatr
technique are in risk for lifelong morbidity. Child Health 2017;53(12):1159–1166

Original Article 117
Current Opinions and Practices of Bariatric

Surgery in Adolescents: A Survey among
Pediatric Surgeons
Yvonne G. M. Roebroek1,2, Suzanne R. Pruijssers1,2, Nicole D. Bouvy1,2 Ernest L. W. E. van Heurn2,4
1 Department of Surgery, Maastricht University Medical Centre, Address for correspondence Yvonne G. M. Roebroek, Department of
Maastricht, The Netherlands Surgery, Maastricht University Medical Centre, P Debyelaan 25, Maastricht
2 Department of Surgery, NUTRIM School for Nutrition, Toxicology 6229HX, The Netherlands (e-mail: [email protected]).
and Metabolism, Universiteit Maastricht, Maastricht, Limburg, The
Netherlands
3 Department of General Surgery, Maastricht University Medical
Centre, Maastricht, The Netherlands
4 Paediatric Surgical Center of Amsterdam, Emma Children’s Hospital,
AMC, VU University Medical Center, Amsterdam, The Netherlands
Eur J Pediatr Surg 2020;30:117–121.
Abstract Introduction Bariatric surgery is performed at increasing rate in severely obese

adolescents who do not respond to conservative treatment. In the United States,
this treatment is generally accepted, yet in Europe, surgeons are more reluctant
because of concerns regarding safety and (long-term) efficacy. We evaluated in which
(European) countries bariatric surgery is allowed and performed, and the opinion of the
members of the European Paediatric Surgeons’ Association (EUPSA) regarding bariatric
surgery in adolescents.
Materials and Methods Information was obtained with an online questionnaire sent
to all EUPSA members.
Results A total of 108 pediatric surgeons (PSs) from 33 countries (of which 25
European) completed the survey. Sixty-two PSs (57.4%) from 22 countries stated that
bariatric surgery in adolescents was allowed in their country. In only 14 countries, the
costs were (partially) covered by health care insurances. Overall, 65.7% of the surgeons
(n ¼ 71) believed bariatric surgery may offer a valuable contribution to obtain
substantial long-term weight loss in severely obese adolescents. Fifty-one (47.2%)
reported that these procedures should be performed by a combination of a bariatric
and a PS, while 20 (18.5%) and 16 (14.8%) preferred that these procedures should be
performed solely by a PS or bariatric surgeon, respectively.
Conclusion Although allowed in most European countries, bariatric surgery in
Keywords adolescents is only practiced on a small scale, and the treatment is often excluded
► pediatrics from regular health care reimbursement. The majority of PSs acknowledge the value of
► severe obesity bariatric surgery, which should be performed by a combination of PS and bariatric
► bariatric surgery surgeon.

These authors contributed equally to this article.
received © 2020 Georg Thieme Verlag KG DOI https://doi.org/

January 3, 2020
118 Current Opinions and Practices of Bariatric Surgery in Adolescents Roebroek et al.
Introduction Participants
All members of the EUPSA were invited for the online survey
In the past four decades, the prevalence of childhood and in October 2018. Responses received between October and
adolescent obesity has risen 10-fold, with currently 6% of November 2018 were eligible for inclusion. No reminders
girls (50 million) and nearly 8% of boys (74 million) being were sent to the invited surgeons.
obese worldwide.1 Although various registries show plateau-
ing of the prevalence curves of pediatric obesity, the number Analysis
of children with severe obesity (body mass index [BMI] > 35 - All completed surveys were used for analysis. Continuous data
kg/m2) is still increasing.2,3 In consequence of this obesity are presented as mean standard deviation. Categorical data
pandemic, many efforts have been made to improve the are presented as number (percentage). If answers on questions
treatment of obesity in the young age group.4,5 Combined about national practice and regulations from surgeons from
lifestyle interventions (CLIs) were regarded as the most one country were contradictory, the score for this country was
optimal treatment for obese and severely obese children classified as “unknown.” For the question “if bariatric surgery
and adolescents for a long time. Although significant short- was performed or not,” the answer was classified as “yes” if one
time reduction of BMI can be obtained by conservative of the surgeons who answered “yes” (1) also clarified that he
treatment, the long-term benefits of CLI programs remain himself performed bariatric surgery in adolescents or (2) if this
small.5–7 “yes” could be confirmed by international scientific publica-
Because of the efficacy of bariatric surgery reported in tions about bariatric surgery in adolescents from this country.
adults, there is considerable interest to utilize this treatment
modality in a younger population. Among both bariatric and
Results
pediatric surgeons (PSs), the trend of pediatric bariatric
surgery is gaining attention and various bariatric procedures Six-hundred eighty-six EUPSA members received an invita-
are already performed in adolescents with severe obesity not tion for the survey in October 2018. Of the 118 responses,
responding to conservative treatment.8 Bariatric surgery is 8.5% (n ¼ 10) were partially answered questionnaires and
considered safe and effective by many to obtain substantial these were not included in the analyses. The 108 complete
weight loss and resolution of associated comorbidities.9,10 responses resulted in a response rate of 15.7%.
Despite the favorable results of bariatric surgery in the
literature, implementation of this treatment shows a large Demographics
variety worldwide. While numbers of severely obese ado- In total, 108 PSs from 25 European and 8 non-European
lescents undergoing bariatric surgery in the United States are countries completed the survey (►Fig. 1). The majority of
increasing, surgeons in Europe seem more reluctant to respondents worked in a university hospital in a European
perform such procedures. country (►Table 1). Twenty-eight PSs (25.9%) reported no
The aim of this study is to assess the opinion and practice specialization within pediatric surgery, while pediatric gas-
of PSs in different European countries to treat severe obesity trointestinal surgery was reported by 47 (43.5%). The mean
in adolescents. Therefore, we conducted an online survey working experience in pediatric surgery was 19.8 12.9 years.
among European Paediatric Surgeons’ Association (EUPSA)
members. We evaluated in which European countries bar- National Practice
iatric surgery is allowed and performed, and the opinion of Of 22 countries, of which 16 European countries, it was
EUPSA members about bariatric surgery in adolescents. stated that bariatric surgery in adolescents was allowed or
not forbidden. Unanimous standards for bariatric surgery in
adolescents were, as reported, only present in three coun-
tries (Portugal, Saudi Arabia, and United States).
Questionnaire Further, differences among countries including the num-
A depersonalized questionnaire was designed using an on- ber of centers, the presence of separate pediatric bariatric
line platform for questionnaires and surveys (Survey Monkey surgical units, and differences in reimbursement of the costs
Inc., San Mateo, California, United States) (Appendix A, are listed in ►Table 2.
available in the online version). The questionnaire consisted The number of clinical practices in countries that perform
of 18 questions in total, obtaining information on national such surgery is reportedly “1 to 2 centers,” n ¼ 9; “1 to 5
regulations and current national practice, and second, centers,” n ¼ 5; “3 to 5 centers,” n ¼ 3; “more than 5 centers,”
addressing the individual areas of knowledge, current prac- n ¼ 3; “unknown,” n ¼ 3. This includes one country in which
tice, and attitudes toward bariatric surgery in adolescents. bariatric surgery in adolescents is not allowed, apart from
Except for two demographic questions, all questions were operations in a trial (The Netherlands).
closed (1 dichotomous, 15 multiple choice). Some closed The costs for bariatric surgery are, in most countries,
questions did allow textual remarks. Conditional branching covered by the health care insurance.
was used to create a customized path based on the respon-
dent’s answers. Using this branching technique, the number Personal Practice and Opinion
of questions answered by a single respondent varied be- Fifteen PSs (13.9%) reported to personally perform bariatric
tween 11 and 18 questions. surgery in children and/or adolescents. An additional 13 PSs
European Journal of Pediatric Surgery Vol. 30 No. 1/2020

Current Opinions and Practices of Bariatric Surgery in Adolescents Roebroek et al. 119
Fig. 1 Distribution of respondents (countries in red) around the world.
(12.0%) reported to work in a practice where bariatric surgery of each of these countries was aware of its existence. The
in adolescents was performed by colleagues, being either PSs poor awareness of national guidelines implies a relevant
(n ¼ 8, 7.4%) or adult bariatric surgeons (n ¼ 5, 4.6%). knowledge gap, possibly leading to a decreased access to
The sleeve gastrectomy was the most commonly used this treatment for severely obese adolescents.
operation (sleeve gastrectomy, n ¼ 21; gastric banding, Based on this survey, we conclude that bariatric surgery in
n ¼ 10; Roux-en-Y gastric bypass, n ¼ 10; and other proce- adolescents in Europe is only practiced on a small scale. A
dures, n ¼ 3). relatively small proportion of the respondents in our survey
Overall, 65.7% of the PSs (n ¼ 71) believed bariatric sur- performs bariatric surgery themselves or works in an institu-
gery may offer a valuable contribution to obtain substantial tion where adolescent bariatric surgery is performed. We
long-term weight loss in severely obese adolescents. Nine PSs would expect that PSs who actually do bariatric surgery in
(8.3%) objected to bariatric surgery in adolescents, and the patients of this age group would be more interested in this
rest of the respondents had no clear opinion. To who should survey, and therefore, have completed the survey more often.
do bariatric surgery in adolescents, 51 PSs (47.2%) were in
favor for a combination of both PS and bariatric surgeon, Table 1 Baseline characteristics of respondents
while 20 (18.5%) and 16 (14.8%) PSs preferred either a PS or a
bariatric surgeon, respectively. Respondents (n ¼ 108)
Country of practice
Discussion European 76.9% (n ¼ 83)
This survey aimed to assess the current national and personal Non-European 23.1% (n ¼ 25)
practices of bariatric surgery in adolescents among PSs in Health care setting
Europe and other countries. Although there is plenty of University hospital 80.6% (n ¼ 87)
literature assessing the feasibility and outcome of this treat-
Regional hospital 11.1% (n ¼ 12)
ment, data to assess the actual clinical practice of bariatric
surgery in different countries are lacking. Private practice 5.6% (n ¼ 6)
Our findings demonstrate that most PSs believe that Other 2.8% (n ¼ 3)
bariatric surgery offers a valuable treatment option for Current position
extreme obesity in adolescents who do not respond to
Consultant 81.5% (n ¼ 88)
conservative treatment. The majority of PSs stated that these
Resident/fellow (in training) 8.3% (n ¼ 9)
procedures can be performed legally; however, only few PSs
were able to confirm the existence of national clinical stand- Other 10.2% (n ¼ 11)
ards and more than 20% of the respondents did not know if Associate professor 2.8% (n ¼ 3)
bariatric surgery was allowed in their country. There are Professor 2.8% (n ¼ 3)
specific guidelines for bariatric surgery in adolescents in
Head of department 4.6% (n ¼ 5)
several countries; only a limited number of the respondents

120 Current Opinions and Practices of Bariatric Surgery in Adolescents Roebroek et al.
Table 2 Practice and regulations for bariatric surgery in adolescents per country
Country Number of Bariatric surgery Number of Separate Bariatric surgery

respondents in adolescents: allowed centers pediatric centers in adolescents:
reimbursement
Austria 2 Yes 1–2 Yes HI
Belgium 4 Yes 3–5 Yes Patient
Bosnia–Herzegovina 1 No – – –
Canada 1 Yes 3–5 Yes Hospital
Chile 2 Yes 1–2 No Hospital
Denmark 3 No – – –
Egypt 1 Yes >5 Yes Not stated
Finland 2 Yes 1–2 Unknown HI
France 4 Yes 1–2/3–5 Yes HI
Germany 6 Unknown Unknown – HI, other (patient)
Greece 5 Yes 1–2 No Patient
Hungary 3 No – – –
Iran 1 Yes Unknown – Combination
Ireland 1 Unknown – No –
Italy 11 Yes 1–2/3–5 Yes/No HI, patient
Latvia 1 No – – –
Luxemburg 1 No – – –
a
Netherlands 6 No 1–2 Yes HI
Pakistan 1 Unknown 1–2 Unknown HI, patient
Poland 5 Yes 1–2/3–5 Yes –
Portugal 2 Yes 1–2 Yes Hospital
Rumania 8 Yes 1–2/3–5 No HI
Russia 1 No – – –
Saudi Arabia 1 Yes Unknown No HI
Serbia 1 Yes 1–2 No HI
Slovakia 1 Yes 1–2 Yes HI
South Africa 2 Unknown Unknown No Unknown
Sweden 10 Yes 1–2 No HI
Spain 8 Yes 1–2/3–5 Yes HI
Switzerland 2 Yes Unknown Unknown HI, combination
Turkey 13 Yes 3–5 Unknown HI, combination
United Kingdom 6 Yes >5 Yes HI
United States 1 Yes >5 Yes HI
Abbreviation: HI, health care insurance.

a
Only allowed in a clinical trial.
However, it is possible that bariatric surgeons do bariatric with supposedly inadequate evidence of its benefits. A study of
operations in this age group without PSs knowing this and Inge et al demonstrated that age younger than 18 years is by far
we have not sent the survey to bariatric surgeons to confirm our the most common provided reason for denial in reimburse-
findings. ment of costs of adolescent bariatric surgery.11
An important factor that may affect variation in bariatric Information about bariatric surgery in adolescents has been
surgical activity among countries is the reimbursement of the collected from 33 different countries, including 25 European
costs for bariatric operations in adolescents. This differs from countries, so that a very good overview of adolescent bariatric
reimbursement in adults as adolescent bariatric surgery is surgical activity in Europe is obtained. As the epidemic of
often not seen as a standard treatment for extreme obesity extreme obesity is also increasing in Europe, the conclusion

Current Opinions and Practices of Bariatric Surgery in Adolescents Roebroek et al. 121
that adolescent bariatric surgery in Europe will grow in the and distribution of the invitation to complete the survey
near future seems justified, although the percentage of se- to the members of EUPSA.
verely obese youth in the United States is still much higher than
in Europe. Another interesting finding is that the vast majority
References
of the PSs who participated in the survey reported that there
1 Abarca-Gómez L, Abdeen ZA, Hamid ZA, et al; NCD Risk Factor
should be a role for the PS to do this type of surgery either Collaboration (NCD-RisC). Worldwide trends in body-mass index,
together with a bariatric surgery or not. The special position of underweight, overweight, and obesity from 1975 to 2016: a
adolescents in whom behavior is developing rapidly and in pooled analysis of 2416 population-based measurement studies
whom feeding and nutritional deficiencies including vitamins, in 128·9 million children, adolescents, and adults. Lancet 2017;
minerals, and trace elements are vital for normal development 390(10113):2627–2642
2 van Dommelen P, Schönbeck Y, van Buuren S, HiraSing RA. Trends
may necessitate a wide view on bariatric operations and an
in a life threatening condition: morbid obesity in dutch, Turkish
important role for the PS. and Moroccan children in The Netherlands. PLoS One 2014;9(04):
The majority of the respondents also recognizes a role for e94299
an adult bariatric surgeon in a future pediatric bariatric 3 Skinner AC, Skelton JA. Prevalence and trends in obesity and
surgical team. This survey can act as an initiation for further severe obesity among children in the United States, 1999-2012.
JAMA Pediatr 2014;168(06):561–566
assessment of various national practices and perhaps collab-
4 Barlow SE; Expert Committee. Expert committee recommenda-
oration of both PS and bariatric surgeon in the different tions regarding the prevention, assessment, and treatment of
European countries. child and adolescent overweight and obesity: summary report.
Pediatrics 2007;120(Suppl 4):S164–S192
5 Oude Luttikhuis H, Baur L, Jansen H, et al. Interventions for
Conclusion treating obesity in children. Cochrane Database Syst Rev 2009;
(01):CD001872
Bariatric surgery in adolescents is allowed in most European
6 Peirson L, Fitzpatrick-Lewis D, Morrison K, Warren R, Usman Ali
countries; however, only few countries have national clinical M, Raina P. Treatment of overweight and obesity in children and
guidelines for this specific patient group. In most countries, youth: a systematic review and meta-analysis. CMAJ Open 2015;3
bariatric surgery in adolescents appears to be practiced only on (01):E35–E46
a small scale, and patients may be excluded from regular health 7 Reinehr T, Holl RW, Wabitsch M. The German Working Group of
care reimbursement. The majority of PSs reported that bariatric Obesity in Childhood and Adolescence (AGA): improving the
quality of care for overweight and obese children in Germany.
surgery is a valuable treatment in adolescents with extreme
Obes Facts 2008;1(01):26–32
obesity in whom other treatment has failed. It seems crucial 8 Chernoguz A, Chwals WJ. Bariatric surgery needs a seat at the
that experiences both from adult bariatric surgeons and PSs are children’s table: bridging the perception and reality of the role of
shared and consensus in treatment can be reached. bariatric surgery in the treatment of obesity in adolescents. Clin
Ther 2018;40(10):1648–1654
9 Durkin N, Desai AP. What is the evidence for paediatric/
Conflict of Interest
adolescent bariatric surgery? Curr Obes Rep 2017;6(03):
Dr. van Heurn reports to be a member of network office of
278–285
the EUPSA society. 10 Beamish AJ, Reinehr T. Should bariatric surgery be performed in
adolescents? Eur J Endocrinol 2017;176(04):D1–D15
Acknowledgments 11 Inge TH, Boyce TW, Lee M, et al. Access to care for adolescents
The authors acknowledge the support of the EUPSA seeking weight loss surgery. Obesity (Silver Spring) 2014;22(12):
2593–2597
Network Office for providing advice on survey design

Editorial 1
Editorial
European Pediatric Surgical Association (EUPSA)

Mikko Pakarinen1 Benno M. Ure2
1 Section of Pediatric Surgery, University of Helsinki, Helsinki
University Central Hospital, Children’s Hospital, Helsinki, Finland
2 Centre of Pediatric Surgery, Hannover Medical School, Hannover,
Germany
Eur J Pediatr Surg 2020;30:1.
The editorial board of the European Journal of Pediatric Surgery van Heurn, Amsterdam; Martin Lacher, Leipzig; Shawn St.
is very pleased to release the European Pediatric Surgical Peter, Kansas City; Udo Rolle, Frankfurt; Juan Tovar, Madrid;
Association (EUPSA) issue of the journal. The issue contains Benno Ure, Hannover; Tomas Wester, Stockholm; and Augusto
papers presented at the 20th European Congress of Pediatric Zani, Toronto.
Surgery which took place in Belgrade, Serbia, between June 12 Fifty-seven manuscripts were submitted and reviewed.
and 16, 2019. Nineteen were finally accepted for publication and will appear
There have been many scientific highlights during the in this issue. Again, these papers represent the high quality of
meeting and numerous of them have been submitted to be scientific work being done by pediatric surgeons all over the
considered for publication in the European Journal of Pediatric world. We appreciate their publication in the European Journal
Surgery. The submitted manuscripts were reviewed by the of Pediatric Surgery.
EUPSA Publication Committee comprising professors Ernest
Address for correspondence © 2020 Georg Thieme Verlag KG DOI https://doi.org/

Benno M. Ure, MD, Centre of Stuttgart · New York 10.1055/s-0040-1702272.
Pediatric Surgery Hannover, ISSN 0939-7248.
Hannover Medical School,
Carl-Neuberg-Straße 1, 30625
Hannover, Germany
(e-mail: [email protected]).
Original Article
Accelerated Intestinal Epithelial Cell Turnover

Correlates with Stimulated BMP Signaling Cascade
following Intestinal Ischemia–Reperfusion in a Rat
Yoav Ben-Shahar1,2 Zaid Abassi1 Hila Kreizman Shefer1 Yulia Pollak3 Udayan Bhattacharya1
Igor Sukhotnik2,3
1 Laboratory of Intestinal Adaptation and Recovery, Technion Israel Address for correspondence Yoav Ben-Shahar, MD, Laboratory of
Institute of Technology Ruth and Bruce Rappaport Faculty of Intestinal Adaptation and Recovery, Technion Israel Institute of
Medicine, Haifa, Haifa, Israel Technology Ruth and Bruce Rappaport Faculty of Medicine, Haifa,
2 Department of Pediatric Surgery, Dana-Dwek Children's Hospital, Tel Haifa 31096, Israel (e-mail: [email protected]).
Aviv Sourasky Medical Center, Tel Aviv, Israel
3 Laboratory of Intestinal Adaptation and Recovery, Tel Aviv University
Sackler Faculty of Medicine, Tel Aviv, Israel
Eur J Pediatr Surg
Abstract Introduction Bone morphogenetic proteins (BMPs) are a family of proteins that
regulate proliferation and differentiation of intestinal epithelial cells. The purpose of
this study was to evaluate the role of BMP signaling following intestinal ischemia–
reperfusion (IR) in a rat model.
Materials and Methods Male Sprague-Dawley rats were divided into four experimen-
tal groups: Sham-24 and Sham-48 rats underwent laparotomy and were sacrificed 24 or
48 hours later, respectively; IR-24 and IR-48 rats underwent occlusion of superior
mesenteric artery and portal vein for 30 minutes followed by 24 or 48 hours of
reperfusion, respectively. Enterocyte proliferation and apoptosis were determined at
sacrifice. BMP-related genes and protein expression were determined using real-time
polymerase chain reaction, Western blot, and immunohistochemistry for 48 hours
followed by IR.
Results IR rats demonstrated a significant increase in BMP2 (twofold increase,
p < 0.05), BMP4 (sevenfold increase), STAT3 (70% increase), BMPR1 (70% increase)
messenger ribonucleic acid levels in jejunum and was accompanied by a significant
Keywords increase in BMP2 and BMP4 protein levels in jejunum (sixfold increase) (Western blot)
► intestinal ischemia– and upward increase in the number of BMP-positive cells (by immunohistochemistry) in
reperfusion jejunal (48% increase) and ileal (56% increase) villi compared with Sham-48 animals.
► bone morphogenetic Elevation in BMP2 and BMP4 levels was associated with increased rates of cell
protein proliferation and increased cell apoptosis.
► stem cells Conclusion Forty-eight hours following intestinal IR in rats, BMP signaling pathway
► proliferation was stimulated. The increase in BMP signaling pathway activity correlates with
► apoptosis accelerated cell turnover.

September 17, 2019
Role of BMP Signaling following Intestinal IR in a Rat Ben-Shahar et al.
Introduction standard chow and had free access to water. Animals were
fasted 24 hours before operation but had free access to water.
The complex damage process known as intestinal ischemia–
reperfusion (IR) injury causes intracellular alterations by Experimental Design
generating radical oxygen species.1 These radicals lead to Rats were divided randomly into four experimental groups.
endothelial damage,2 increase permeability of the epithelium,3 Group I (Sham-24) underwent laparotomy in which the
and result in dysfunction of the gut barrier. Furthermore, superior mesenteric artery and portal vein were identified
inflammatory-mediated cells infiltrating through intestinal and isolated but were not occluded. Group II (IR-24) under-
wall release cytokines giving rise to systemic inflammatory went laparotomy in which both superior mesenteric artery
response and finally multiple organ failure and death.4,5 Cur- and portal vein were occluded for 20 minutes. Group III
rently, apoptosis that is an active process occurs during reper- (Sham-48) underwent laparotomy the same as the
fusion, is known to have an important key role in cell death first group followed by 48 hours of reperfusion. Group IV
process following IR injury.6,7 (IR-48) underwent occlusion of both superior mesenteric
Intestinal cell turnover is affected by the balance between artery and portal vein followed by 48 hours of reperfusion.
production and differentiation of enterocytes and death by All the animals were sacrificed by injection of pentobarbital
apoptosis. Interactions between the epithelium and underly- (45 mg/kg) intraperitoneally and open pneumothorax 24 or
ing mesenchymal stroma regulate intestinal cell fate. Mature 48 hours after operation.
intestinal epithelium is generated by proliferation of stem cells
situated deeply in intestinal crypts. These cells migrate to the Surgical Procedure
lumen and differentiate to form the villi epithelium.8 The main Rats were anaesthetized by injection of ketamine (90 mg/kg)
signaling pathways involved in the regulation of stem cells and xylazine (15 mg/kg) intraperitoneally after an overnight
activity are Wnt/b-catenin, hedgehog, bone morphogenetic fast. An abdominal incision was made in midline. In the sham
protein (BMP), and notch.9 Growing evidence suggests that groups, both superior mesenteric artery and portal vein were
hedgehog and BMP signaling pathway are important mediator isolated but were not occluded. In the IR groups, both
of these interactions.10,11 superior mesenteric artery and portal vein were occluded
The BMPs are part of a family of growth factors known as by applying atraumatic microvascular clamps. During
transforming growth factor-β that regulate proliferation 20 minutes of ischemic period, the abdominal wall incision
and differentiation of intestinal epithelial cells.12,13 More was covered with sterile pad to prevent fluid and heat loss.
than 30 familiar BMPs classified into groups such as BMP2/4 Next, the abdominal wall was uncovered and the clamp was
group, BMP5/6/7/8 group known as osteogenic protein-1 removed, and the ischemic bowel was returned into the
group, and BMP9/10 group.14 Recent evidence suggests that abdomen. Before closure, the abdominal cavity was washed
BMP2 is expressed in the epithelium and crypt cells of with an intraperitoneal injection of 3-mL saline 0.9%. The
human and mouse colon and BMP4 is a canonical Wnt abdominal wall was sutured with a running suture of 3/0
target gene.15 There are two BMP receptor types known as Vicryl (Ethicon Corporation, New Jersey, United States) in
BMP receptor 1 (BMPR1) (1a or 1b) and BMPR2. BMP ligands two layers. The rats were fasted for 6 hours following opera-
bind to the receptor and lead to phosphorylation of type I tion, but were allowed free access to water. After 24 hours or
receptor and BMP-specific SMADs (SMAD1, SMAD5, and 48 hours, animals were anesthetized with an intraperitoneal
SMAD8) subsequently. Next, the phosphorylated SMADs injection of pentobarbital (75 mg/kg) and then were sacri-
form a complex with SMAD4 and move into the nucleus ficed by open pneumothorax. The small intestine was taken
where it acts as a transcriptional enhancer by binding to out, excised quickly, and rinsed with cold isotonic saline.
other deoxyribonucleic acid (DNA) binding proteins.16 Little Then, the bowel was divided into two segments: proximal
is known about the role of BMP signaling cascade in jejunum and terminal ileum. Each segment was weighed and
stimulation of intestinal epithelial cells turnover, especially the intestinal mucosa was scraped using a microspatula,
following intestinal IR. collected and weighed. Histologic sections were made
The purpose of this study was to evaluate the role of BMP from jejunal and ileal segments of each group.
signaling in stimulating stem cell activity following intestinal
IR in a rat model. Intestinal Histology
Small bowel segments were fixed in 4% buffered formalin for
24 hours and adapted into standard paraffin blocks. Tissue
slices of five-micron thick were stained with hematoxylin–
Animals eosin. Two expert investigators blinded to the experimental
All animal experiments were performed under the standards groups evaluated the damage to intestinal mucosa by using
for care and use of animals as stated in the Guide for the Use the Park’s criteria.17 A score scaled at 0 to 8 represents the
and Care of Laboratory Animals (Bruce Rappaport Faculty of severity: 0—normal mucosa, 1—subepithelial space at villus
Medicine, Technion-Israel Institute of Technology, Haifa, tip, 2—more extended subepithelial space, 3—epithelial
Israel). Sprague-Dawley male rats (weighing 250–300 g) lifting along villus sides, 4—denuded villi, 5—loss of villus
were held in pairs on cycles of 12-hour day and night for tissue, 6—crypt layer infarction, 7—transmucosal infarction,
3 days at 21°C before the experiment. The rats were fed with 8—transmural infarction.

Immunohistochemistry After electrophoresis, proteins were transferred to PVDF

Standard 5-bromodeoxyuridine labeling reagent (Zymed Labo- membrane and probed with anti-BMP 2/4 antibody (1:500
ratory, Inc., California, United States) was injected to the animals dilution, sc-137087), and antitubulin antibody (1:1000 dilution,
at a dose of 1 mL per 100 g body weight 90 minute before sc-53029). Horseradish peroxidase-conjugated secondary anti-
sacrifice to detect enterocyte proliferation. Five-micrometer body was purchased from Jackson Immuno Research Laborato-
slices (5 µm) embedded with paraffin were deparaffinized ries Inc. (West Grove, Pennsylvania, United States) and an
with xylene and rehydrated with graded alcohol. The slices enhanced chemiluminescent substrate from Biological Indus-
were then stained with a biotinylated monoclonal anti-BrdU tries (Kibbutz Beth Ha-Emek, Israel).
antibody system using BrdU Staining Kit (Zymed Laboratory,
Inc, California, United States). The ratio of crypt cells staining Statistical Analysis
positively for BrdU per 10 crypts determined the index of The data are expressed as the mean standard error of mean.
proliferation. Identification of apoptotic cells was done by Kruskal–Wallis test was used for statistical analysis of param-
immunohistochemistry for caspase-3 (Caspase-3 cleaved con- eters of adaptation, enterocyte proliferation, and apoptosis.
centrated polyclonal antibody; dilution 1:100; Biocare Medical, Then, post hoc test was done for multiple comparisons with
Walnut Greek, California, United States) using a combination of p-values of less than 0.05 considered statistically significant.
streptavidin–biotin–peroxidase method and microwave
antigen retrieval on formalin-fixed, paraffin-embedded tissues
Results
according to the manufacturers’ protocols. The detection of
apoptotic cells was made by counting the number of stained Intestinal Mucosal Parameters
cells in at least 10 villi in areas without necrosis for each group. IR-24 rats (Group B) have shown a significant decrease in
The apoptotic index was determined as the number of apoptotic bowel weight in jejunum (15% decrease, p < 0.05), mucosal
cells per 10 villi. To detect BMP signaling pathway activity, weight in jejunum (25% decrease, p < 0.05), and ileum (21%
immunohistochemistry for BMP2 and BMP4 (BMP2/4 mono- decrease, p < 0.05), as well as a trend toward decrease in
clonal antibody, dilution 1:100, sc-137087) was performed for villus height and crypt depth in jejunum and ileum; however,
identification of positive cells using a combination of the this trend was not statistically significant (►Table 1). IR-24
streptavidin–biotin–peroxidase method and microwave anti- rats demonstrated a significant increase in Park's injury
gen retrieval on formalin-fixed, paraffin-embedded tissues score in jejunum (2.5-fold increase, p < 0.05) and ileum
according to the manufacturer’s protocols. All positive cell (twofold increase, p < 0.05) suggesting intestinal damage.
measurements were performed by a qualified pathologist IR-48 rats (Group D) demonstrated less intestinal damage in
blinded as to the source of intestinal tissue. both jejunum and ileum that was accompanied by a signifi-
cant increase in mucosal weight and villus height in ileum
Real-Time PCR compared with IR-24 animals, suggesting intestinal recovery.
Trizol (Invitrogen) reagent was used to isolate ribonucleic
acid (RNA) according to manufacturer’s protocol. Spectro- Intestinal Epithelial Cell Turnover (Proliferation and
photometry of 260/280 nm was used to quantify the Apoptosis)
extracted RNA. Thereafter, 500 ng of total RNA was converted We have shown previously18 that 24 hours following intestinal
into complementary DNA (cDNA) by reverse transcriptase IR injury, rats demonstrated a strong decrease in cell prolifer-
(PrimeScript RT reagent Kit TaKaRa, Japan). Next, cDNA was ation rates in jejunum and ileum, compared with control
amplified by PCR-Thermal Cycler (2720 Thermal Cycler, ABI, animals. Current experiment has shown that 48 hours after
Israel). Gene expression of BMP2, BMP4, STAT3, BMPR1, and IR, experimental rats demonstrated a small but significant
BMPR2 messenger RNA (mRNA) was determined by quanti- increase in intestinal epithelial cell proliferation in jejunum
tative real-time polymerase chain reaction (PCR) ABI-PRISM (12% increase, p < 0.05) and ileum (13% increase, p < 0.05)
7000 (applied Biosystems, Foster City, California, United compared with sham animals (►Fig. 1). Significantly greater
States) on cDNA samples using Cyber Green Master Mix numbers of apoptotic cells appeared in the villi of jejunum
(Takara, Japan) with the exception of template and primers. (threefold increase, p < 0.05) and ileum (threefold increase,
p < 0.05) in IR rats compared with sham animals (►Fig. 1).
Western Blotting However, a total number of apoptotic cells was lower in IR-48
Protein extraction from rat jejunal and ileal tissue samples was compared with IR-24 animals (Group B).
performed by homogenizing in radio immunoprecipitation
assay lysis buffer containing 50 mM Tris–HCl (pH 7.4), BMP Signaling Related Genes (Real-Time PCR)
150 mM NaCl, 1% NP-40, 2 mM ethylenediaminetetraacetic BMP signaling-related genes expression was determined by
acid, supplemented with a cocktail of protease (Roche Diagnos- real-time PCR (►Fig. 2). As expected, IR rats demonstrated a
tic) and phosphatase cocktail inhibitors (Sigma). The homoge- significant increase in BMP2 (twofold increase, p < 0.05),
nate was centrifuged at 7500 rpm at 4°C for 15 minutes and the BMP4 (sevenfold increase, p < 0.05), STAT3 (70% increase,
supernatant was collected. The protein concentrations were p < 0.05), BMPR1 (70% increase, p < 0.05) mRNA levels in
determined by Bradford reagent according to manufacturer’s jejunum compared with sham animals. In ileum, BMP-related
instructions. Samples containing equal amounts of total protein gene expression was not significantly different between IR and
(30 µg) were resolved by SDS–PAGE under reducing conditions. sham animals.

Table 1 Intestinal mucosal parameters
Parameters Sham-24 IR-24 Sham-48 IR-48

Park's score
Jejunum 0.6 0.2 1.5 0.2a 0.4 0.2 0.9 0.3a,b
Ileum 0.9 0.2 2 0.3a 0.8 0.2 1.5 0.3a
Bowel weight (mg/cm/100 g BW)
Jejunum 26 2 22 1a 22 1 20 2
Ileum 21 1 19 1 21 2 18 1
Mucosal weight (mg/cm/100 g BW)
Jejunum 10 1 7.5 0.3a 10 1 7.6 0.4a
a
Ileum 8 0.5 6.3 0.4 9 0.4 7.5 0.7a,b
Villus height (µm)
Jejunum 615 69 551 70 679 38 639 51
Ileum 420 35 380 39 549 27 496 22a,b
Crypt depth (µm)
Jejunum 176 9 180 10 163 10 171 12
Ileum 168 11 155 6 157 15 164 11
Abbreviations: BW, body weight; IR, ischemia–reperfusion; SE, standard error.

Note: Value are mean SE.
a
p < 0.05 IR versus Sham.
b
p < 0.05 IR-48 versus IR-24.
Fig. 1 Effect of intestinal IR on intestinal cell proliferation and apoptosis. Values are mean standard error of mean. IR, ischemia–reperfusion.

p < 0.05 IR rats versus sham rats.

Fig. 2 Effect of intestinal IR on BMP signaling related genes. Gene expression of BMP2, BMP4, STAT3, BMPR1, and BMPR2 mRNA was determined
by quantitative real-time polymerase chain reaction. Values are mean standard error of mean. BMPR1, bone morphogenetic protein receptor 1;
IR, ischemia–reperfusion. p < 0.05 IR rats versus sham rats.
Western Blot Discussion

A significant upregulation of BMP signaling-related genes in
ischemic animals was accompanied by a significant increase in Intestinal epithelial cell turnover after intestinal IR is an
BMP2 and BMP4 protein levels in jejunum (sixfold increase, incredibly complex process that relies heavily on evolutionarily
p < 0.05) compared with sham animals as well as in a trend conserved signaling pathways to provide crucial cell commu-
toward increase in BMP2/4 protein levels in ileum; however, nication. Typically, secreted signaling proteins such as Wnts,
this trend was not statistically significant (►Fig. 3). BMPs, and Hedgehogs released by one cell population will
trigger concentration-dependent responses in other cells
Immunohistochemistry located some distance away. BMPs are members of the
In damaged intestine, transient expression of the BMP2/4 was TGF-β superfamily, and more than 30 different BMP isoforms
observed in pericryptal mesenchymal cells and intestinal have been identified in mammals and Drosophila. BMPs were
epithelial stem cells, which may contribute to this originally described as osteoinductive cytokines that enhance
required second signal. Intestinal IR injury resulted in a bone and cartilage formation when injected into mice and
marked increase in the number of BMP2 and BMP4 positive were found to play an important role in the regulation of
cells in jejunal (48% increase, p < 0.05) and ileal (56% increase, embryonic development and cellular homeostasis as well,
p < 0.05) villi compared with sham animals (►Fig. 4). The including the regulation of proliferation, differentiation, apo-
number of BMP2/4 positive cells in crypts was not different ptosis, and remodeling of the extracellular matrix.12,19 The role
between IR and sham animals. of BMP signaling cascade in stimulation of intestinal stem cell
Fig. 3 Effect of intestinal IR on BMP2 and BMP4 protein expression. Western blot was used to determine protein levels. Values are
mean standard error of mean. BMP2, bone morphogenetic protein 2; IR, ischemia–reperfusion. p < 0.05 IR rats versus sham rats.

Fig. 4 Immunohistochemistry for BMP 2 and BMP 4 (BMP 2/4 monoclonal antibody, dilution 1:100, sc-137087) was performed for the
identification of positive cells using a combination of the streptavidin–biotin–peroxidase method and microwave antigen retrieval on formalin-
fixed, paraffin-embedded tissues according to the manufacturer’s protocols. Values are mean standard error of mean. BMP2, bone
morphogenetic protein 2; IR, ischemia–reperfusion. p < 0.05 IR rats versus sham rats.
activity after intestinal IR has never been previously investi- However, 48 hours after IR injury, intestinal epithelial cell
gated. Radhakrishnan et al have recently demonstrated that proliferation was increased, suggesting tissue repair.8 This
pretreatment with BMP7 mimics ischemia preconditioning increase in cell proliferation was accompanied by increased
following intestinal IR injury in the intestine and liver in a rat cell apoptosis suggesting accelerated intestinal epithelial cell
model.20 The authors conclude that BMP7 protected against turnover probably due to activated intestinal stem cell activity.
intestinal IR induced intestinal and liver injury and that BMP7 We next investigated the possible role of BMP signaling in
may be a more logical surrogate to ischemia preconditioning in activating intestinal stem cell activity after intestinal IR
the prevention of injury in the setting of intestinal IR. injury. BMP pathway-related gene expression was deter-
The purpose of this study was to evaluate the role of BMP mined by real-time PCR. We have shown that 48 hours
signaling during intestinal recovery following intestinal IR in a following IR event, accelerated intestinal epithelial cell turn-
rat model. Similar to our previous experiments,8,18 we have over was accompanied by a strong increase in BMP2 (twofold
demonstrated in the current study that 24 hours following increase), BMP4 (sevenfold increase), STAT3 (70% increase),
intestinal IR injury experimental rats demonstrated a signifi- BMPR1 (70% increase) mRNA levels in jejunum compared
cant increase in intestinal injury score that was accompanied with sham animals. In ileum, BMP-related gene expression
by an intestinal mucosal hypoplasia. This is evident from a was not significantly different between IR and sham animals.
significant decrease in bowel weight and mucosal weight as This increase in BMP-related gene expression was coin-
well as a trend toward decrease in villus height and crypt cided with a significant upregulation of BMP2 and BMP4
depth. IR-48 rats (Group D) demonstrated less intestinal protein levels in jejunum (sixfold increase) compared with
damage in both jejunum and ileum that was accompanied sham animals as well as in a trend toward increase in BMP2/4
by a significant increase in mucosal weight and villus height in protein levels in ileum; however, this trend was not statisti-
ileum compared with IR-24 animals, suggesting intestinal cally significant. Immunostaining for BMP2/4 in damaged
recovery. The increase in intestinal epithelial cell proliferation intestine has shown a transient expression of the BMP2/4 in
in jejunum and ileum (in IR-48 rats compared with sham pericryptal mesenchymal cells and intestinal epithelial stem
animals) may be responsible for this positive effect. Analysis of cells, which may contribute to this required second signal.
epithelial proliferation after IR injury was performed using Intestinal IR injury resulted in a marked increase in the
BrdU incorporation as a marker, demonstrated elevated DNA number of BMP2 and BMP4 positive cells in jejunal (48%
synthesis in the epithelium of the entire small intestine, and increase) and ileal (56% increase) villi compared with sham
enhanced rates of cell proliferation. We have previously dem- animals. The number of BMP2/4 positive cells in crypts was
onstrated that 24 hours after intestinal IR injury, rats demon- not different between IR and sham animals.
strated a significant increase in Park's injury score, decrease in Transient expression of the BMP2/4 has been observed in
cell proliferation rates, and increased rates of cell apoptosis pericryptal mesenchymal cells and intestinal epithelial stem
suggesting intestinal damage rather than intestinal recovery.18 cells, which may contribute to this required second signal. The

correlation between BMP and MAPK signaling and cell prolif- following intestinal ischemia-reperfusion in a rat. Pediatr Surg Int
eration has been reported by other investigators. Ge et al have 2019;35(02):255–261
demonstrated that extracellular matrix and BMP regulation of 9 de Santa Barbara P, van den Brink GR, Roberts DJ. Development
and differentiation of the intestinal epithelium. Cell Mol Life Sci
transcriptional activity in osteoblasts is predominantly medi-
2003;60(07):1322–1332
ated by extracellular signal–regulated kinases (ERKs).21 Kosin- 10 Roberts DJ, Smith DM, Goff DJ, Tabin CJ. Epithelial-mesenchymal
ski et al have recently investigated the gene expression profiles signaling during the regionalization of the chick gut. Develop-
of the human colon by comparing the gene expression pattern ment 1998;125(15):2791–2801
between the top and basal crypt compartments.22 BMP antag- 11 Bitgood MJ, McMahon AP. Hedgehog and Bmp genes are coex-
pressed at many diverse sites of cell-cell interaction in the mouse
onists, including gremlin 1, expressed by intestinal subepithe-
embryo. Dev Biol 1995;172(01):126–138
lial myofibroblasts (ISEMFs) and smooth muscle cells in turn
12 Hogan BL. Bone morphogenetic proteins: multifunctional regu-
create an opposite gradient to antagonize BMPs, thus main- lators of vertebrate development. Genes Dev 1996;10(13):
taining Wnt activity at the crypt base. This gradient then 1580–1594
creates an environment that promotes stem cell self-renewal 13 Bandyopadhyay A, Yadav PS, Prashar P. BMP signaling in develop-
and expansion at the crypt base region. ment and diseases: a pharmacological perspective. Biochem
Pharmacol 2013;85(07):857–864
In conclusion, in a rat model of intestinal IR, 48 hours after
14 Shroyer NF, Wong MH. BMP signaling in the intestine: cross-talk is
IR event enhanced epithelial cell turnover was associated with key. Gastroenterology 2007;133(03):1035–1038
activated BMP signaling that coincided with previously de- 15 Hardwick JCH, Van Den Brink GR, Bleuming SA, et al. Bone
scribed inhibited. This activated BMP pathway in the peri- morphogenetic protein 2 is expressed by, and acts upon, mature
cryptal mesenchymal cells and intestinal epithelial stem cells, epithelial cells in the colon. Gastroenterology 2004;126(01):
thereby resulting in increased crypt cell proliferation. 111–121
16 Sukhotnik I, Berkowitz D, Dorfman T, et al. The role of the BMP
signaling cascade in regulation of stem cell activity following
Conflict of Interest massive small bowel resection in a rat. Pediatr Surg Int 2016;32
None declared. (02):169–174
17 Park PO, Haglund U, Bulkley GB, Fält K. The sequence of develop-
ment of intestinal tissue injury after strangulation ischemia and
References reperfusion. Surgery 1990;107(05):574–580
1 Carden DL, Granger DN. Pathophysiology of ischaemia-reperfu- 18 Sukhotnik I, Brod V, Lurie M, et al. The effect of 100% oxygen on
sion injury. J Pathol 2000;190(03):255–266 intestinal preservation and recovery following ischemia-reperfu-
2 Wolin MS. Interactions of oxidants with vascular signaling sys- sion injury in rats. Crit Care Med 2009;37(03):1054–1061
tems. Arterioscler Thromb Vasc Biol 2000;20(06):1430–1442 19 Sampath TK, Maliakal JC, Hauschka PV, et al. Recombinant human
3 Mangino JE, Kotadia B, Mangino MJ. Characterization of hypo- osteogenic protein-1 (hOP-1) induces new bone formation in vivo
thermic intestinal ischemia-reperfusion injury in dogs. Effects of with a specific activity comparable with natural bovine osteo-
glycine. Transplantation 1996;62(02):173–178 genic protein and stimulates osteoblast proliferation and differ-
4 Schoenberg MH, Poch B, Younes M, et al. Involvement of neutrophils entiation in vitro. J Biol Chem 1992;267(28):20352–20362
in postischaemic damage to the small intestine. Gut 1991;32(08): 20 Radhakrishnan RS, Radhakrishnan GL, Radhakrishnan HR, et al.
905–912 Pretreatment with bone morphogenetic protein-7 (BMP-7)
5 Yamamoto S, Tanabe M, Wakabayashi G, Shimazu M, Matsumoto mimics ischemia preconditioning following intestinal ischemia/
K, Kitajima M. The role of tumor necrosis factor-alpha and reperfusion injury in the intestine and liver. Shock 2008;30(05):
interleukin-1beta in ischemia-reperfusion injury of the rat small 532–536
intestine. J Surg Res 2001;99(01):134–141 21 Ge C, Yang Q, Zhao G, Yu H, Kirkwood KL, Franceschi RT. Inter-
6 Ikeda H, Suzuki Y, Suzuki M, et al. Apoptosis is a major mode of cell actions between extracellular signal-regulated kinase 1/2 and
death caused by ischaemia and ischaemia/reperfusion injury to p38 MAP kinase pathways in the control of RUNX2 phosphoryla-
the rat intestinal epithelium. Gut 1998;42(04):530–537 tion and transcriptional activity. J Bone Miner Res 2012;27(03):
7 Noda T, Iwakiri R, Fujimoto K, Matsuo S, Aw TY. Programmed cell 538–551
death induced by ischemia-reperfusion in rat intestinal mucosa. 22 Kosinski C, Li VS, Chan AS, et al. Gene expression patterns of
Am J Physiol 1998;274(02):G270–G276 human colon tops and basal crypts and BMP antagonists as
8 Ben-Shahar Y, Pollak Y, Bitterman A, Coran AG, Bejar IN, Sukhotnik I. intestinal stem cell niche factors. Proc Natl Acad Sci U S A 2007;
Sonic hedgehog signaling controls gut epithelium homeostasis 104(39):15418–15423

Original Article
Mitochondrial DNA: A Biomarker of Disease

Severity in Necrotizing Enterocolitis
Edoardo Bindi1,2 Bo Li1 Haitao Zhou1 Maarten Janssen Lok1 Mashriq Alganabi1
Rossella Angotti2 Agostino Pierro1
1 Division of General and Thoracic Surgery, Translational Medicine Address for correspondence Agostino Pierro, OBE, MD, FRCS, FAAP,
Program, The Hospital for Sick Children, Toronto, Ontario, Canada Division of General and Thoracic Surgery, Robert M. Filler Professor of
2 Division of Pediatric Surgery, Department of Medical Sciences, Surgery, University of Toronto, The Hospital for Sick Children,
Surgical Sciences and Neurosciences, Hospital of “Santa Maria alle 1526-555 University Avenue, Toronto, ON M5G 18, Canada
Scotte,” Siena, Italy (e-mail: [email protected]).
Eur J Pediatr Surg
Abstract Introduction There is a need to develop sensitive markers to diagnose or monitor the
severity of intestinal damage in necrotizing enterocolitis (NEC). Mitochondrial deox-
yribonucleic acid (mtDNA) is increased in the intestine and blood of adults in response
to intestinal ischemia and can trigger secondary organ damage. We hypothesize that
mtDNA is increased during experimental NEC and that mtDNA levels are correlated to
the degree of intestinal injury.
Materials and Methods NEC was induced in C57BL/6 mice (n ¼ 18) (approval: 44032)
by gavage feeding with hyperosmolar formula, hypoxia, and lipopolysaccharide
administration from postnatal day (P) 5 to 9. Breastfed pups served as control
(n ¼ 15). Blood was collected by cardiac puncture and terminal ileum was harvested
on P9. Reverse transcription quatitative polymerase chain reaction was used to
measure mtDNA (markers COX3, CYTB, ND1) and inflammatory cytokines (interleukin
6 [IL-6] and tumor necrosis factor-α[TNF-α]) in blood and ileum. Intestinal injury was
scored blindly by four investigators and classified as no/minor injury (score 0 or 1) or
NEC (score 2).
Results mtDNA is significantly increased in gut and blood of NEC mice (p < 0.05).
Furthermore, mtDNA increases in intestine and blood proportionally to the degree of
Keywords intestinal injury as indicated by a positive correlation with histological scoring and
► necrotizing inflammation (r ¼ 0.6; p < 0.05) (expression of IL-6 and TNF-α).
enterocolitis Conclusion Following NEC intestinal injury, mtDNA is released from the intestine into
► mitochondrial DNA circulation. The blood level of mtDNA is related to the degree of intestinal injury.
► disease severity mtDNA can be a novel marker of intestinal injury and can be useful for monitoring the
► biomarker progression of NEC.
Introduction infants where it can reach around 20 to 50%. Those who survive
are often confronted with severe morbidity due to NEC such as
Necrotizing enterocolitis (NEC) is a devastating intestinal short bowel syndrome and neurodevelopmental delay.1
disease that mainly affects preterm infants. NEC is character- Diagnosing NEC can be challenging since the disease often
ized by high mortality specifically in very low birth weight progresses quickly, and a delay in diagnosis may increase

August 20, 2019
Biomarker of Disease Severity in Necrotizing Enterocolitis Bindi et al.
morbidity and/or mortality due to NEC. Biomarkers should Histology

facilitate the diagnosis of NEC and monitoring its progres- Ileum was fixated in 4% formalin, embedded in paraffin, and
sion. Until now, a variety of biomarkers have been studied in sectioned in 5 µm. Hematoxylin/eosin staining was performed.
blood and urine during NEC, such as calprotectin, claudin, Intestinal damage was scored (0–3) blindly by four indepen-
and intestinal fatty acid binding protein (i-FABP), but none of dent investigators and classified as no injury (0), minor injury
them offers reliable and reproducible results.2 (1), or NEC-like injury ( 2). NEC was diagnosed on the basis of
Mitochondrial deoxyribonucleic acid (mtDNA) is a circular mucosal edema, disarrangement of villus enterocytes, severe
double-stranded molecule, contained in the mitochondria, villus core separation, and epithelial sloughing of the villi.8
which contains 37 genes coding for proteins involved in the
electronic respiratory chain. Recent studies have shown that Reverse Transcription Quatitative Polymerase Chain
mtDNA, released from necrotic cells, acts like a damage- Reaction
associated molecular pattern (DAMP) triggering inflammation Cell-free plasma was obtained from blood samples by centri-
in tissues and remote organ injury through activation of the fugation (3,500 revolutions per minute, 4°C, 10 minutes). DNA
Toll-like receptor 9 (TLR-9) receptor.3,4 It has been shown that extraction from ileum and plasma was done using TRizol
in the intestine, the release of mtDNA following ischemic- Reagent (Thermo Fisher Scientific, Inc., Illinois, United States).
reperfusion injury results in gut barrier dysfunction.5 Gu et al Real-time polymerase chain reaction (PCR) was performed
demonstrated that high plasma levels of mtDNA are strongly with Advanced qPCR Master Mix (Wisent Inc., Quebec, Canada)
associated with complications and worse prognosis in patients and CFX384 Real-Time System (Bio-Rad Laboratories, Inc).
with trauma.6 These authors reported that mtDNA could be an mtDNA levels in ileum and plasma were measured by quanti-
independent predictor of systemic inflammatory response fication of three different sequences corresponding to the
syndrome.6 COX3, CYTB, and ND1 mitochondrial genomic regions. Gene
In this study, we aimed to evaluate whether mtDNA could expression analysis of inflammatory cytokines interleukin 6
be a useful and reliable biomarker for the severity and (IL-6) and tumor necrosis factor α (TNF-α) was performed for
progression of experimental NEC. both intestine and plasma.9 The expression of each gene was
normalized to the expression of the housekeeping gene GAPDH
(IL-6, TNF-α) and 18s rRNA (COX3, CYTB, ND1). Sequence of
primer in each gene is shown in ►Table 1.
Experimental Model of NEC
Ethical approval was obtained from the ethical review board of Statistics
The Hospital for Sick Children (#44032). C57BL/6 mice pups GraphPad Prism6 (GraphPad Software Inc., San Diego,
were randomly divided into control and NEC groups on post- California, United States) was used for statistical analyses.
natal day 5 (P5). Controls (n ¼ 15) stayed with the mothers and Data were analyzed using Student’s t-test. p-Values of < 0.05
were breastfed. NEC was induced in pups (n ¼ 18) from P5 to P9, were considered significant.
based on a previous animal model.7,8 NEC induction comprises
several steps: (1) maternal separation on P5; (2) 10 minutes of
Results
hypoxia 4 times a day before each feeding; (3) hyperosmolar
formula feeding by gavage 4 times a day (0.05 mL/g body Intestinal Injury
weight) (7.5 g Similac Lower Iron [Abbott Laboratories, Ltd., NEC was associated with a severe mucosal injury in the ileum
Saint-Laurent, QC, Canada] þ 34.5 mL Esbilac Puppy Milk Re- as indicated by histological analysis (►Fig. 1A) and increased
placer [PetAg, Inc., Hampshire, Illinois, United States]); and (4) expression of proinflammatory cytokines. Histologic score
lipopolysaccharide (4 mg/kg) enteral administration together indicated a significant difference (p < 0.0001) between con-
with the first gavage feeding on P6 and P7. On P9, a blood trol and NEC (►Fig. 1B). None of the control animals showed
sample was collected by cardiac puncture and then pups were NEC-like injury, whereas 14 (77%) NEC animals showed NEC-
sacrificed by decapitation. The distal ileum was harvested. like injury. The intestine of NEC animals showed higher
Table 1 List of primers used for RT-qPCR
Gene Forward sequence (5–3) Reverse sequence (5–3)

COX3 CGTGAAGGAAACTACCCAGG CGCTCAGAAGAATCCTGCAA
CYTB TGAGGGGGCTTCTCAGTAGA CTGTTTCGTGGAGGAAGAGG
ND1 GGATCCGAGCATCTTATCCA GGTGGTACTCCCTCTGTAAA
IL6 CCAATTTCCATTGCTCTCCT ACCACAGTGAGGAATGTCCA
TNFα TTCCGAATTCACTGGAGCCTCGAA TGCACCTCAGGGAAGAATCTGGAA
GAPDH TGAAGCAGGCATCTGAGGG CGAAGGTGGAAGAGTGGGAG
18s rRNA GTAACCCGTTGAACCCCATT CCATCCAATCGGTAGTAGCG
Abbreviation: RT-qPCR, reverse transcription quatitative polymerase chain reaction.

Fig. 1 Intestinal injury. (A) hematoxylin and eosin (HE) staining of control and necrotizing enterocolitis (NEC) ileum. (B) Injury score. (C, D)
Cytokines levels in the ileum.
expression of IL-6 (p < 0.05) and TNF-α (p < 0.05) compared Taken together, our results showed that the severity of
with control (►Fig. 1C and D). Epithelial injury and inflam- mucosal injury correlated significantly with the level of
mation indicate the presence of significant intestinal injury mtDNA both in the ileum (r ¼ 0.6; p < 0.05) and blood
during NEC. (r ¼ 0.6; p < 0.05) (►Fig. 4A and B).
Mitochondrial Deoxyribonucleic Acid Levels

Discussion
Levels of mtDNA in the ileum of NEC-induced pups were
higher compared with controls (COX3 p < 0.05, CYTB This study indicates that mtDNA is increased in response to
p < 0.05, ND1 p < 0.05) (►Fig. 2A). Similarly, mtDNA levels the intestinal injury seen in NEC. The mtDNA elevation can be
were higher in the plasma of NEC-induced pups compared detected in circulating blood, raising the possibility of using
with controls (COX3 p < 0.05, CYTB p < 0.05, ND1 p < 0.05) this marker for future NEC diagnosis and monitoring.
(►Fig. 2B). Multicenter studies reported that: (1) the median time
between initial clinical concern and confirmed medical NEC
Intestinal Injury Correlates with mtDNA Levels is 31 hours and to surgery is 57 hours10; and (2) the
Severe intestinal injury in NEC, indicated by histological majority of patients that develop advanced NEC do not
score 2, was associated with high mtDNA levels in the present with a fulminant course, resulting in disease pro-
gut. Greater severity of mucosal injury in the intestine gression.11 Therefore, it is important to discover new bio-
correlated with higher levels of mtDNA (p < 0.05) markers which are able to diagnose NEC early and to
(►Fig. 3A). The same correlation was also observed in monitor its progression. An early diagnosis and treatment
circulating blood: higher the intestinal epithelial injury, can improve the outcome of NEC and minimize the risk of
higher the mtDNA (p < 0.05) (►Fig. 3B). progressing to multiorgan failure.
Fig. 2 Mitochondrial deoxyribonucleic acid (mtDNA) levels. (A) mtDNA levels in the ileum. (B) mtDNA levels in the blood.

Fig. 3 Intestinal injury is associated with mitochondrial deoxyribonucleic acid (mtDNA) levels. In necrotizing enterocolitis (NEC) the more
severe the histologic injury in the intestine, the higher are the levels of mtDNA in the ileum (A) and in the blood (B).
Fig. 4 Intestinal injury correlates with mitochondrial deoxyribonucleic acid (mtDNA) levels. Severity of mucosal injury correlated significantly
with the levels of mtDNA in both the ileum (A) and blood (B).
Levels of cytokines have been demonstrated to be higher extracellular space. Cell contents are normally packaged into
in both intestine and plasma of patients with NEC compared apoptotic bodies that become phagocytosed without ever
with those with feeding intolerance or other noninfectious leaking directly into the extracellular space, and therefore,
diseases. However, this difference is not significant when without triggering immune responses.16 On the other hand,
compared with patients with septic ileus or other intestinal under pathologic conditions, such as ischemia-reperfusion
surgical diseases, suggesting that cytokines cannot be used injury, fragments of mtDNA are released from necrotic cells.
as specific biomarkers.12 In these conditions, the mtDNA acts as a DAMP and stim-
Other markers related to intestinal pathophysiology, such ulates TLR-9. In this way upon release into the circulation,
as i-FABP and calprotectin, have also been investigated. In mtDNA can trigger inflammatory responses in the tissue of
2010, Evennett et al showed a significant increase of i-FABP origin, as well as in other organs. Gu et al demonstrated how
in infants with NEC requiring surgery.13 In a systematic intratracheal administration of exogenous mtDNA could
review in 2016, calprotectin was found to be high in stools induce severe lung injury in mice through TLR-9 activation.17
of NEC patients. However, studies included in the review had Similarly, in a rat model of heart ischemia-reperfusion injury,
a high variability in the sensitivity and specificity of calpro- the tail injection of mtDNA enhanced myocardial necrosis.18
tectin.14 These issues together with limited prospective These studies have increased interest on mtDNA in the
studies, render most NEC biomarkers published to date of clinical setting and highlighted the usefulness of mtDNA as
limited clinical utility. potential biomarker of various diseases (trauma, tumors,
Recently, there has been considerable interest in mtDNA, intestinal ischemia).
particularly circulating mtDNA, as a potential marker for Five recent studies3–6,19 have been performed in adults
various diseases. Mitochondria are not only the power- admitted to intensive care units following severe traumatic
house of the cell, but are also involved in calcium homeo- injury. These studies indicate that levels of mtDNA are
stasis, cell death, and thermoregulation. The mitochondrial higher in these patients compared with healthy controls,
genome is composed of a double-stranded circular DNA and that mtDNA has a positive correlation with the severity
molecule which has a lot of similarities with the bacterial of injury and with markers of systemic inflammatory
genome. This could be explained by the theory that eu- response. In addition, among patients who had a severe
karyotic cells originated from the fusion of proteobacteria trauma, those with higher levels of mtDNA had the worse
and archaebacteria, with the latter giving rise to the outcome. Although these studies have limitations due to
mitochondria.15 small sample size and variability in the extent of injury,
Under physiologic conditions, cell turnover in the form of they report encouraging results for the usefulness of mtDNA
apoptosis does not result in release of free mtDNA into the as clinical biomarker.

NEC is a disease that presents several risk factors, but its 2 Goldstein GP, Sylvester KG. Biomarker discovery and utility in
pathophysiology is not completely understood. However, one necrotizing enterocolitis. Clin Perinatol 2019;46(01):1–17
3 Hauser CJ, Sursal T, Rodriguez EK, Appleton PT, Zhang Q, Itagaki K.
of the main pathologic mechanisms of NEC is ischemia-reper-
Mitochondrial damage associated molecular patterns from fem-
fusion injury. NEC involves derangement of intestinal micro-
oral reamings activate neutrophils through formyl peptide recep-
circulation that leads to poor blood flow, which may induce an tors and P44/42 MAP kinase. J Orthop Trauma 2010;24(09):
inflammatory cascade, resulting in intestinal injury.20 534–538
In this study, we hypothesized that in NEC, the intestinal 4 Zhang Q, Itagaki K, Hauser CJ. Mitochondrial DNA is released by
injury leads to release of mtDNA from the necrotic intestinal shock and activates neutrophils via p38 map kinase. Shock 2010;34
(01):55–59
epithelial cells. The increased levels of mtDNA could enhance
5 Hu Q, Ren H, Ren J, et al. Released mitochondrial DNA following
the inflammatory injury trough activation of TLR-9 receptors
intestinal ischemia reperfusion induces the inflammatory re-
while mtDNA released in the circulation triggers injury in sponse and gut barrier dysfunction. Sci Rep 2018;8(01):7350
remote organs. 6 Gu X, Yao Y, Wu G, Lv T, Luo L, Song Y. The plasma mitochon-
In our model of experimental NEC, the intestine drial DNA is an independent predictor for post-traumatic
presented severe mucosal injury, as demonstrated by his- systemic inflammatory response syndrome. PLoS One 2013;
8(08):e72834
tological changes and increased inflammation (elevated
7 Barlow B, Santulli TV, Heird WC, Pitt J, Blanc WA, Schullinger JN.
expression of IL-6 and TNF-α). We demonstrated that the An experimental study of acute neonatal enterocolitis–the im-
NEC-induced ischemic injury was associated with increased portance of breast milk. J Pediatr Surg 1974;9(05):587–595
intestinal and circulating levels of mtDNA. In addition, we 8 Zani A, Cordischi L, Cananzi M, et al. Assessment of a neonatal rat
demonstrated a positive relationship between the degree of model of necrotizing enterocolitis. Eur J Pediatr Surg 2008;18(06):
mucosal injury and levels of mtDNA: greater the histologi- 423–426
9 Miyake H, Li B, Lee C, et al. Liver damage, proliferation, and
cal damage, higher the levels of mtDNA in the circulating
progenitor cell markers in experimental necrotizing enterocolitis.
blood and in the intestine. These findings support the J Pediatr Surg 2018;53(05):909–913
hypothesis that mtDNA could be considered not only as 10 Sylvester KG, Ling XB, Liu GY, et al. A novel urine peptide
an index of injury in the intestine but also a contributing biomarker-based algorithm for the prognosis of necrotising en-
factor in the pathogenesis of NEC. terocolitis in human infants. Gut 2014;63(08):1284–1292
11 Ji J, Ling XB, Zhao Y, et al. A data-driven algorithm integrating
These initial results encourage further clinical research to
clinical and laboratory features for the diagnosis and prognosis of
define the role of mtDNA as a biomarker of NEC, but also to
necrotizing enterocolitis. PLoS One 2014;9(02):e89860
understand its involvement in remote organ damage associ- 12 Gephart SM, Gordon PV, Penn AH, et al. Changing the paradigm of
ated with this disease. defining, detecting, and diagnosing NEC: perspectives on Bell’s stages
and biomarkers for NEC. Semin Pediatr Surg 2018;27(01):3–10
Authors’ Contributions 13 Evennett NJ, Hall NJ, Pierro A, Eaton S. Urinary intestinal fatty
acid-binding protein concentration predicts extent of disease in
E.B., B.L., and A.P. designed experiments; E.B. performed
necrotizing enterocolitis. J Pediatr Surg 2010;45(04):735–740
experiments; E.B. wrote the manuscript; and A.P. provid- 14 Pergialiotis V, Konstantopoulos P, Karampetsou N, et al. Calpro-
ed advice and supervision; all the authors reviewed and tectin levels in necrotizing enterocolitis: a systematic review of
revised the manuscript. the literature. Inflamm Res 2016;65(11):847–852
15 Skulachev VP. Mitochondrial physiology and pathology; concepts
Conflict of Interest of programmed death of organelles, cells and organisms. Mol
Aspects Med 1999;20(03):139–184
None declared.
16 Thurairajah K, Briggs GD, Balogh ZJ. The source of cell-free
mitochondrial DNA in trauma and potential therapeutic strate-
Acknowledgments gies. Eur J Trauma Emerg Surg 2018;44(03):325–334
The authors would like to acknowledge the Laboratory 17 Gu X, Wu G, Yao Y, et al. Intratracheal administration of mito-
Animal Services of The Hospital for Sick Children for their chondrial DNA directly provokes lung inflammation through the
technical expertise and assistance with animal work. A.P. TLR9-p38 MAPK pathway. Free Radic Biol Med 2015;83:149–158
18 Xie L, Liu S, Cheng J, Wang L, Liu J, Gong J. Exogenous administration of
is the recipient of a Canadian Institutes of Health Research
mitochondrial DNA promotes ischemia reperfusion injury via TLR9-
(CIHR) Foundation Grant 353857 and the Robert M. Filler p38 MAPK pathway. Regul Toxicol Pharmacol 2017;89:148–154
Chair of Surgery, The Hospital for Sick Children. 19 Zhang J, Chen X, Liu Z, et al. Association between plasma mitochon-
drial DNA and sterile systemic inflammatory response syndrome in
patients with acute blunt traumatic injury. Int J Clin Exp Med 2017;
References 10(02):3254–3262
1 Horbar JD, Edwards EM, Greenberg LT, et al. Variation in perfor- 20 Alganabi M, Lee C, Bindi E, Li B, Pierro A. Recent advances in
mance of neonatal intensive care units in the United States. JAMA understanding necrotizing enterocolitis. F1000 Res 2019;8:8
Pediatr 2017;171(03):e164396

Original Article
Clinical Factors in Trunk Capillary Malformations in the

Neonate: When to Suspect Other Associated
Malformations? A Case–Control Study
Carlos Delgado-Miguel1 Miriam Vicente1 Antonio Jesus Serrano-Muñoz1 Miriam Miguel-Ferrero1
Mercedes Diaz1 Paloma Triana1 Juan Carlos López-Gutiérrez2
1 Department of Pediatric Surgery, Hospital Universitario La Paz, Address for correspondence Carlos Delgado-Miguel, MD,
Madrid, Spain Department of Pediatric Surgery, Hospital Universitario La Paz,
2 Division of Vascular Anomalies, Department of Pediatric Surgery, La Paseo de la Castellana, 261, Madrid 28046, Spain
Paz Children’s Hospital, Madrid, Spain (e-mail: [email protected]).
Eur J Pediatr Surg
Abstract Introduction Capillary malformations (CMs) can be sporadic or syndromic, in

association with other underlying venous malformation (VM) or lymphatic malforma-
tion (LM). The objective of this study is to describe the clinical patterns in the neonate
that allow us to differentiate sporadic CMs from those associated with other vascular
malformations.
Materials and Methods A case–control study was performed in neonates with CM
located in the trunk, followed at our institution between 2008 and 2018. The patients
were divided into two groups: group A (cases: CM associated with VM or LM) and group
B (controls: sporadic CM without associated malformations). Demographic and clinical
variables collected in the clinical history were evaluated (color, location, multifocality,
bilaterality, position regarding the vascular axis, and involvement of the midline).
Results Thirty-eight patients were included (18 cases and 20 controls) without
differences in gender and age. In group A, the totality of patients presented CM
with uniform color and lateral location (p < 0.001). In this group, bilateral and
multifocal involvements were lower than in group B, without significant differences
Keywords between both groups. The distribution of CMs in group A was always parallel to the
► capillary vascular axis and the midline was always respected, without observing these character-
malformations istics in the group B (p < 0.001).
► vascular Conclusion The presence of a CM in the trunk of a neonate with uniform color, lateral
malformations location, parallel position to the vascular axis, and absence of involvement of the
► neonates midline should make us suspect other underlying vascular malformations, which
► case–control study should be studied with complementary tests.
Paloma Triana’s ORCID is https://orcid.org/0000-0002-8176-

6010.

July 4, 2019
Clinical Factors in Trunk Capillary Malformations in the Neonate Delgado-Miguel et al.
Introduction Materials and Methods
Capillary malformations (CMs), revealed as cutaneous stains, A case–control study was performed in neonates with CMs
are a frequent complaint in pediatric consultations.1 Its located in the trunk, followed at our institution in the past
origin is found in the capillary vessels mainly of the skin, 10 years (2008–2018). Patients were divided into two groups
where capillaries and postcapillary venules are dilated, based on the study factor analyzed: presence of associated
although they can also involve other organs or systems.2 underlying LMs or venous malformations (VMs). Group A
Although most CMs are isolated cases, they can some- (cases) was formed by patients with CMs who presented
times be associated with other underlying malformations, associated LMs or VMs, while group B (controls) was formed
either vascular, other soft tissues, or even skeletal, giving by patients with isolated CMs without associated vascular
rise in some patients to syndromes with repeated and malformations. To determine the presence of underlying
recognizable phenotypic patterns that suggest a common associated malformations, all patients underwent comple-
physiopathogenic mechanism. Some genetic bases of these mentary tests at the time of diagnosis, Doppler ultrasound as
vascular spots and associated syndromes have been the first step in all cases, and, on occasion, supplemented
recently identified. Most of them are related to activating with magnetic resonance imaging (MRI) in patients in which
mutations of the Ras-MAP kinase or the phosphatidylino- ultrasound did not determine the extent of underlying
sitol-3-kinase pathway. This is the case of megalence- associated malformations.
phaly-capillary malformation syndrome and congenital The descriptive analysis was performed through the
lipomatous overgrowth, vascular malformations, epider- review of the medical records and the visualization of
mal nevi, spinal/skeletal anomalies/scoliosis syndrome, the photographs taken in the consultation of our unit, with
both caused by different activating mutations of the the prior consent of the legal tutors, with the aim of
PIK3CA gene.3 describing and characterizing the CMs presented in both
The importance of these skin spots lies in their ability to groups. We analyzed demographic variables (age and sex),
be potential markers or warning signals that allow an early personal history (diseases and previous surgeries, concomi-
diagnosis of possible associated malformations. This diag- tant medical pathology, and treatment), clinical data of the
nosis can be complex due to the overlap of clinical pre- CM (color, location, multifocality, bilaterality, position
sentations and its low frequency.4 However, in situations in regarding the vascular axis, and involvement of the midline),
which the finding of a CM is suspected, either by its and associated malformations and symptoms.
location, extension, or association with other alterations, All patients who presented a malformation of capillary
the use of complementary studies will help achieve the origin located in the anatomical area of the trunk whose
correct diagnosis and implementation of the adequate medical history was available in the hospital files and whose
therapeutic measures.5 lesion had been documented by photographs were included.
Therefore, early diagnosis and correct identification of Patients who had been diagnosed or included within a
these malformations are a priority. Through the use of syndrome or illness that explained the presence of said
appropriate classifications and knowledge of well-defined lesion were excluded.
syndromes, we can optimize the clinical management of For the statistical analysis, the continuous variables were
these children, since many of them will need specific inter- expressed as mean and standard deviation. To check whether
vention and long-term follow-up.6 the variables were normally distributed, the Kolmogorov–
From the review of the literature, it can be suspected that Smirnoff’s and Shapiro–Wilk’s tests were used. For the
CMs in newborns do not appear randomly and their charac- continuous variables normally distributed, the Student’s
teristics point toward the presence of associated anomalies t-test of independent samples was used, and to analyze the
and genetic mutations. The clinical morphology of CMs has continuous data not normally distributed, the Mann–Whit-
been associated with a higher or lower risk of other vascular ney’s U test was used. The discrete variables were expressed
abnormalities. In a study of patients diagnosed with Klippel– as frequency and percentage, and were analyzed by the chi-
Trénaunay’s syndrome, the finding of “geographic” CMs square test or Fisher’s test when the first one could not be
(those CMs characteristically well delineated, with clear applied. Odds ratios were calculated with 95% confidence
borders and a deep red color) was associated with an intervals. All statistical calculations were performed with
underlying lymphatic malformation (LM), as well as a greater two tails and the statistical significance was established with
risk of subsequent complications.7 a value of p < 0.05. The data were collected in Microsoft Excel
In this context, we hypothesized that the characteristics of software version 2010 and analyzed with SPSS Statistic,
these malformations, when located in the trunk wall, would version 22 (Chicago, Illinois, United States).
also be able to alert about the presence of underlying
anomalies and be related to certain genetic alterations.
Results
The main objective of this study is to analyze the clinical
patterns in the neonate that may allow us to differentiate Thirty-eight patients were included (18 cases and 20 con-
isolated CMs from those associated with other vascular trols), all of them diagnosed of CMs in the wall of the trunk, in
malformations, and to assess their possible role as a predic- all cases present at the time of birth, without differences in
tive factor for the diagnosis of other underlying disorders. gender and age. In group A, the totality of patients presented

Table 1 Demographic and morphological characteristics
Capillary malformations associated with Sporadic capillary p–Value OR (95% CI)

venous/lymphatic malformations (n ¼ 18) malformations (n ¼ 20)
Age (mo) 2.6 0.5 3.2 0.6 0.84 –
Gender
Women 8 (44%) 9 (45%) 0.97 1.02 (0.28–3.68)
Men 10 (56%) 11 (55%)
Bilaterality 2 (11%) 4 (20%) 0.45 0.50 (0.08–3.13)
Multifocality 6 (33%) 8 (40%) 0.67 0.75 (0.20–2.83)
Absence of midline 18 (100%) 3 (15%) <0.001 –
involvement
Location lateral 18 (100%) 0 <0.001 –
Parallel position to 18 (100%) 0 <0.001 –
the vascular axis
Uniform color 18 (100%) 0 <0.001 –
Abbreviations: CI, confidence interval; OR, odds ratio.
CMs with uniform color and lateral location, without observ- shows the demographic and morphological characteristics of
ing these characteristics in group B (p < 0.001). However, in both groups.
this group, 20% of patients had bilateral CMs and 40% of them Regarding the associated malformations in group A, all of
had multifocal involvement, which was higher than in group them were lymphatic and venous types, with a variable size
A (11 and 33%, respectively), without significant differences that was not related to the extension of the CMs. These were
between both groups (p ¼ 0.45 for bilaterality and p ¼ 0.67 always in the same anatomical area as the CMs, located in
for multifocality). The distribution of CMs in group A was the subcutaneous tissue and underlying it although some-
always parallel to the vascular axis and the midline times exceeding its limits. ►Fig. 1 shows the images of
was always respected, while none of these characteristics patients who presented CMs associated with VM and LM.
was described in patients in group B (p < 0.001). ►Table 1 In all the patients, the associated alterations were discovered
Fig. 1 (A–H) Photographs of patients who presented capillary malformations of the trunk wall with other associated malformations. Note the
laterality of the lesion, its arrangement parallel to the vascular axis and its color pattern.

simultaneously with the diagnosis of the CM, through Dopp- Discussion

ler ultrasound, and the study was later extended with the
realization of an MRI, to better characterize the lesion and CMs are the most common vascular abnormalities in chil-
delimit it in extension and depth. dren occurring in 0.3 to 2.1% of newborns.1 Although most
In all cases, the associated malformation had a venous CMs found in normal examinations will be isolated vascular
component, minor and asymptomatic, and a lymphatic com- malformations, it is essential to consider other vascular or
ponent, predominant and symptomatic. The LMs that pre- skeletal abnormalities when they show specific character-
sented the group of cases are defined as an abnormal istics in their morphology7 or when they are found at specific
proliferation of the lymphatic vessels. They were located on anatomic sites.8,9 Its importance resides, therefore, in their
the trunk in all cases, underlying the CMs. Eleven of the patients ability to be red flags for other disorders. In those situations
(61%) presented LMs with a size similar to the associated CM, in which the finding of CMs is highly suspicious, performing
while in seven of them, the LM surpassed the limits of the CM. early complementary analyses will allow for a diagnosis and
Due to the importance of the LM and its clinical impact, mainly appropriate therapeutic management.5
pain, all patients underwent surgical resection. Four patients Considering the results of our study, it is possible to deduce
were operated on three occasions, 3 on two occasions, and 11 that in patients presenting with CMs, the characteristics of the
on one occasion. In two patients, it was also necessary to skin lesions can guide us toward the diagnosis of associated
establish a treatment with rapamycin (mammalian target of malformations, with great diagnostic relevance and therapeutic
rapamycin inhibitor) to reduce the symptomatology of the significance. In our cases, these characteristics were specified in
residual LM after the surgical intervention. No action was the form, location and color of the CM, being highly indicative of
necessary regarding the venous component. The control presenting underlying VMs or LMs when presented unilaterally,
patients in group B did not present any symptomatology and without crossing the midline and with a homogeneous color
no treatment was needed. It is remarkable the absence of any pattern. The presence of concomitant symptoms, generally
repetitive or recurrent pattern in the CMs presented in the absent in isolated CMs, should also alert us to a possible
control group, which are shown in ►Fig. 2. associated malformation.
In neither of the two groups was family aggregation In CMs, the onset age is usually within the first month, but
observed, with no history of vascular malformations in the in the LMs and VMs, the onset age is often after the first
parents or first-degree relatives, and none of the patients was month and, depending on the location, they may go unno-
related to another individual in the study. Nor was there any ticed for years before their diagnosis. The reason may be
relevant medical or surgical background common in any of partly because the lesions of CMs is reddish or dark red
the patients of both groups. patches on the surface of skin, and easy to be noticed by
Fig. 2 (A–H) Photographs of patients who presented capillary malformations of the trunk wall without other associated malformations. Note
the absence of any repetitive pattern in the lesions of these patients.

scrupulous parents. As for other vascular malformations, the this field, for the repercussion that an incorrect diagnosis and
lesions usually the same color as skin, and take place deep a consequent inadequate management of the associated
under the skin, caused an overlook in the early stages.10 pathology it can entail for the patient.
Many authors agree that the complementary tests to be
performed should consist, in first instance, of performing a
Conclusion
Doppler ultrasound in the anatomical area of the lesion to
reveal the extent of interaction with surrounding structures The presence of a CM in the trunk of a neonate with uniform
as well as for procedural planning, being able to complete the color, lateral location, parallel position to the vascular axis,
study with an MRI, if deemed appropriate, having been and absence of involvement of the midline should make us
demonstrated usefulness in the diagnosis of VM and LM suspect other underlying vascular malformation.
associated.8,11 However, the application of other tests such as
lymphangiography, phlebography, or angiography is more Conflict of Interest
controversial, and the complexity of its performance and the None declared.
risks involved limit its use to complex cases.
It is advisable, therefore, when CMs of the trunk that
References
suggest the suspicion of associated malformations are found,
1 Jacobs AH, Walton RG. The incidence of birthmarks in the neonate.
either by its location (unilateral), its distribution (without Pediatrics 1976;58(02):218–222
crossing the midline), or its color pattern (homogeneous), 2 Rozas-Muñoz E, Frieden IJ, Roé E, Puig L, Baselga E. Vascular stains:
that the neonatologist or the pediatrician performs comple- proposal for a clinical classification to improve diagnosis and
mentary tests and informs the parents of the possibility that, management. Pediatr Dermatol 2016;33(06):570–584
eventually, the spot may be a sign of other vascular alter- 3 Boon LM, Ballieux F, Vikkula M. Pathogenesis of vascular anoma-
lies. Clin Plast Surg 2011;38(01):7–19
ations in the child.
4 Uller W, Fishman SJ, Alomari AI. Overgrowth syndromes with
After the diagnosis of underlying vascular malformations, complex vascular anomalies. Semin Pediatr Surg 2014;23(04):
referral to a specialized center in the management of vascu- 208–215
lar anomalies should be considered. Combined vascular 5 López-Gutiérrez JC, Redondo P, Ivars M. Fingertip capillary mal-
malformations are challenging and require a multidisciplin- formation and associated disorders: report of 9 cases. Pediatrics
2017;140(01):e20162967
ary team for its treatment.11
6 Dasgupta R, Fishman SJ. ISSVA classification. Semin Pediatr Surg
In VMs, the size and extent of the lesion determine the 2014;23(04):158–161
prescribed therapy, with small well-localized VMs often 7 Maari C, Frieden IJ. Klippel-Trénaunay syndrome: the importance
treated successfully in a single session with a single modality of “geographic stains” in identifying lymphatic disease and risk of
(e.g., resection and sclerotherapy). In contrast, large exten- complications. J Am Acad Dermatol 2004;51(03):391–398
sive VMs that involve surroundings structures are rarely 8 Waelchli R, Aylett SE, Robinson K, Chong WK, Martinez AE, Kinsler
VA. New vascular classification of port-wine stains: improving
curable and may require multiple treatment modalities.12,13
prediction of Sturge-Weber risk. Br J Dermatol 2014;171(04):
In our patients, the venous component was minimal and 861–867
asymptomatic, so an observation attitude was decided. 9 Sillard L, Léauté-Labreze C, Mazereeuw-Hautier J, et al. Medial
Regarding the lymphatic component, some authors defend fronto-facial capillary malformations. J Pediatr 2011;158(05):
the role of laser therapy or sclerotherapy in the treatment of 836–841
complex lymphatic anomalies, although they recognize a 10 Yang B, Li L, Zhang LX, Sun YJ, Ma L. Clinical characteristics and
treatment options of infantile vascular anomalies. Medicine
temporary solution in them.14 In our patients, due to the
(Baltimore) 2015;94(40):e1717
extension and importance of the lymphatic anomaly that 11 Azizkhan RG. Complex vascular anomalies. Pediatr Surg Int 2013;
they presented, surgical resection was considered more 29(10):1023–1038
appropriate, which was performed in all patients. 12 Behravesh S, Yakes W, Gupta N, et al. Venous malformations:
In relation to the application of these notions in clinical clinical diagnosis and treatment. Cardiovasc Diagn Ther 2016;6
(06):557–569
practice, the neonatologist and the pediatrician should be
13 Hage AN, Chick JFB, Srinivasa RN, et al. Treatment of venous
familiar with the recognition of patterns of appearance of
malformations: the data, where we are, and how it is done. Tech
CMs whose presence can alert of associated alterations and Vasc Interv Radiol 2018;21(02):45–54
start up the pertinent explorations in each case. It is, 14 Trenor CC III, Chaudry G. Complex lymphatic anomalies. Semin
therefore, of vital importance the training of specialists in Pediatr Surg 2014;23(04):186–190

Original Article
The Impact of Anti-Tumor Necrosis Factor Alpha

Therapy on Postoperative Complications in
Pediatric Crohn’s Disease
Vojtech Dotlacil1 Jiri Bronsky2 Ondrej Hradsky2 Barbora Frybova1 Stepan Coufal3
Richard Skaba Michal Rygl1
1
1 Department of Pediatric Surgery, 2nd Faculty of Medicine, Charles Address for correspondence Vojtech Dotlacil, MD, Department of
University in Prague, University Hospital Motol, Prague, Czech Republic Pediatric Surgery, Charles University, V Uvalu 84, Praha 5 150 06,
2 Department of Pediatrics, 2nd Faculty of Medicine, Charles University in Czech Republic (e-mail: [email protected]).
Prague, University Hospital Motol, Prague, Czech Republic
3 Laboratory of Cellular and Molecular Immunology, Institute of Microbiology
of The Czech Academy of Sciences, v.v.i. Prague, Czech Republic
Eur J Pediatr Surg
Abstract Introduction The incidence of Crohn’s disease (CD) within the pediatric population is
increasing worldwide. Despite a growing number of these patients receiving anti-
tumor necrosis factor α therapy (anti-TNF-α), one-third of them still require surgery.
There is limited data as to whether anti-TNF-α influences postoperative complications.
We evaluated postoperative complications in patients who were or were not exposed
to anti-TNF-α therapy in our institutional cohort.
Materials and Methods A retrospective review of CD patients who underwent abdomi-
nal surgery between September 2013 and September 2018 was performed. The patients
were divided into two groups based on whether they were treated with anti-TNF-α within
90 days before surgery. Thirty-day postoperative complications were assessed using
Clavien–Dindo classification (D-C); this examination included surgical site infections
(SSIs), stoma complications, intra-abdominal septic complications, non-SSIs, bleeding,
ileus, readmission rate, and return to the operating room. Mann–Whitney U-test, Fisher’s
exact test, and multivariate logistic regression analyses were used for statistical analysis.
Results Sixty-five patients (41 males) with a median age of 16 years (range: 7–19) at
Keywords the time of operation were identified. The most common surgery was ileocecal
► pediatric surgery resection in 49 (75%) patients. Forty-three (66.2%) patients were treated with anti-
► Crohn’s disease TNF-α preoperatively. Seven patients (11%) experienced postoperative complications.
► anti-tumor necrosis There was no statistically significant difference in postoperative complication in
factor α patients who did or did not receive anti-TNF-α before surgery (D-C minor 2.3% vs.
► postoperative 4.6%, p ¼ 1; D-C major 7% vs. 9.1%, p ¼ 1).
complications Conclusion The use of anti-TNF-α in pediatric CD patients within the 90 days prior to
► inflammatory bowel their abdominal surgery was not associated with an increased risk of 30-day postoper-
disease ative complications.

August 20, 2019
Anti-TNF-α and Postoperative Complications in Pediatric CD Patients Dotlacil et al.
Introduction groups based on their exposure to this agent within the

90 days prior to surgery.
Crohn’s disease (CD) is a chronic, relapsing, and remitting
inflammatory condition that occurs in adults and children. Its Clinical Variables of Interest
incidence in the pediatric population is increasing worldwide. The data set included age, gender, body mass index, history
The phenotype is often characterized by extensive inflamma- of diabetes mellitus, and medical history (corticosteroids or
tion and an aggressive and progressive disease course.1,2 immunomodulators—azathioprine, mercaptopurine, metho-
During the past decade, there have been significant trexate, and biologics). In terms of biologics, only anti-TNF-α
advances in CD treatment with the addition of biologic was included (type, dosage, treatment start, timing, treat-
therapies such as infliximab, a chimeric anti-tumor necrosis ment duration, and last administration before surgery).
factor α (anti-TNF-α) antibody, which revolutionized the CD Disease duration, character, and history (disease behavior
therapeutic regime.3 Unfortunately, the overall effect of anti- according to the Paris classification), reasons for surgery
TNF-α is limited; one-third of patients experience primary (acute complications that required emergency surgery—
nonresponse4–6 and another third have secondary loss of gastrointestinal bleeding, intestinal perforation or obstruc-
response.7,8 Despite the obvious benefit of anti-TNF-α tion), timing of the operation, type of procedure performed,
agents, one-third of pediatric patients still develop compli- whether the surgical procedure was restorative (resection
cations such as fistula, stricture, and obstruction and require with anastomosis), length of hospital stay, 30-day postoper-
surgery within 5 years of their initial diagnosis.9 ative complications, and follow-up were analyzed. Compli-
The high risk of surgery in the pediatric CD population and cations were identified during the hospital stay and/or
the introduction of anti-TNF-α demand the need to examine within the postoperative follow-up.
the effect of biologics on postoperative complications. Rela-
tively robust analysis are already available in adults,10–12 but Outcomes
data on pediatric patients are still missing. Our study retro- The primary outcome was 30-day postoperative complications
spectively analyzed a cohort of pediatric CD patients treated at with respect to preoperative anti-TNF-α treatment. Complica-
the authors’ institution to compare the rate of 30-day postop- tions were graded using the Clavien–Dindo classification
erative complications, readmission and return to operating (D-C).15 They were characterized specifically as surgical site
room (ROR) based on preoperative exposure to anti-TNF-α infections (SSIs; superficial, deep, and organ/space), stoma
agents. We hypothesized that treatment with anti-TNF-α is complications, intra-abdominal septic complications (anasto-
associated with increased postoperative complications. motic leakage, intra-abdominal abscess, or enterocutaneous
fistula), non-SSIs, ileus (bowel obstruction with radiographic
confirmation and requirement of a nasogastric tube for decom-
pression or operative intervention), readmission, and ROR.
Patient Population
Electronic medical records were reviewed to identify all pedi- Statistical Analysis
atric CD patients ( 19 years of age) who underwent surgery at a The continuous variables were tested for normality using the
single tertiary center between September 2013 and September D’Agostino–Pearson normality test. The differences between
2018. All patients were diagnosed with CD based on Porto anti-TNF-α and non-anti-TNF-α groups were analyzed with
criteria13 and were treated according to the European Crohn’s the Mann–Whitney test (continuous variables) or by Fisher’s
and Colitis Organisation/European Society for Paediatric exact test (dichotomous variables). Continuous variables are
Gastroenterology, Hepatology and Nutrition (ESPGHAN) guide- presented as median (range or interquartile range [IQR]), and
lines.14 One surgical team performed all patient operations. The dichotomous variables as the number of cases (percentages),
type of procedure was chosen by the surgeon. Open surgery was and p-values of < 0.05 were considered statistically signifi-
performed via vertical midline laparotomy, and three-port cant. To account for potential confounding effects, we per-
access was used in laparoscopic procedures (one umbilical, formed multivariate logistic regression analyses.
one left epigastric, and one in the left iliac fossa). Mobilization
and resection were achieved using an advanced electrocautery Ethical Commission
device, followed by anteposition of the ileocecal region in front This study was approved by the authors’ institutional ethical
of the abdominal wall for both approaches. In all cases, end-to- committee.
end anastomosis was performed in one layer with interrupted
Vicryl (Johnson & Johnson) stitches. Patients with isolated
Results
perianal surgeries, abdominal surgeries without suturing the
intestine, and patients who received biologics other than anti- Sixty-five patients (41 males, 63%) were included in the
TNF-α were excluded. study. The median age was 16.4 years (IQR: 14.6–17.1) at
the time of operation and 13 years (IQR: 10.9–15.7) at the
Study Design time of diagnosis. None of the patients were smokers or had a
A retrospective chart review (analysis of medical records) history of diabetes mellitus. The remaining demographic
was performed to evaluate postoperative complications. The data, disease behavior, and comparison between groups
patients were divided into anti-TNF-α and non-anti-TNF-α are summarized in ►Table 1.

Table 1 Demographic variables for the entire cohort
Variable Overall; n ¼ 65 Anti-TNF-α; n ¼ 43 No anti-TNF-α, n ¼ 22 p-Value

(66.15%) (33.85%)
n (%) n (%) n (%)
Female 24 (37) 15 (35) 9 (41) 0.7866
Age at surgery, years (median, IQR) 16.4 (14.6–17.1) 15.7 (14.8–17.0) 16.8 (16.1–17.2) 0.0673
Age at diagnosis, years (median, IQR) 13 (10.9–15.7) 12.7 (9.3–14.7) 15 (12.6–16.4) 0.0199
Disease duration, months (median, IQR) 28 (9–44) 32 (15–44) 14.5 (2.8–42.8) 0.0417
Diabetes mellitus 0 0 0 1.0
BMI, kg/m2 (median, IQR) 18.3 (16.3–20.5) 17.8 (16.1–20.0) 19.1 (16.5–20.1) 0.5102
BMI
< 18 31 (47.7) 22 (51.1) 9 (40.9) 0.6003
18–24 30 (46.2) 18 (41.9) 12 (54.6) 0.432
25–29 3 (4.6) 2 (4.7) 1 (4.5) 1.0
30–34 0 (0.0) 0 0 (0.0) 1.0
35 1 (1.5) 1 (2.3) 0 (0.0) 1.0
Disease behavior
B1 (nonstricturing/penetrating) 13 (20) 12 (27.9) 1 (4.5) 0.0458
B2 (stricturing) 42 (64.4) 27 (62.8) 15 (68.2) 0.7866
B3 (penetrating) 10 (15.4) 4 (9.3) 6 (27.3) 0.0756
Perianal modifier 10 (15.4) 9 (20.1) 1 (4.55) 0.1449
Abbreviations: BMI, body mass index; IQR, interquartile range; TNF-α, tumor necrosis factor α.
Note: Data presented as numbers (%).
Forty-three (66.2%) patients received preoperative anti- group. There was no statistically significant difference between
TNF-α (infliximab 60.5%; adalimumab 39.5%) with a median the two groups according to the D-C classification (D-C minor
of 17 months total anti-TNF-α use prior to their operation 2.3% vs. 4.6%, p ¼ 1; D-C major 7% vs. 9.1%, p ¼ 1). The anti-TNF-
(IQR: 6–35). The administration of concomitant medication α treatment did not influence the postoperative complication
in these groups is summarized in ►Table 2. The last dose of rate even when compared with other variables (►Table 4).
anti-TNF-α in all patients was received within the 90-day There was one ROR and one case of ileus/bowel obstruction in
period before surgery: median of 26 days (IQR: 14–47). Anti- both groups. None of the patients experienced non-SSIs
TNF-α dosages were within the standard maintenance inter- (urinary tract infection, pneumonia, or bacteremia). In the
val in 25 (58%) patients. observed anti-TNF-α cohort, there was one case with superfi-
There were no significant differences in the type of cial wound infection that was successfully managed on an
surgery performed between the two groups. The most outpatient basis and one case of ileostomy bleeding that was
frequent surgery was ileocecal resection in 49 (75%) patients, treated conservatively. The absolute 30-day mortality rate was
and 61 (94%) surgical procedures were restorative. All anas- zero. Median follow-up after surgery was 23 months (IQR:
tomoses were fashioned as end-to-end. The median operat- 7–41). All postoperative data are summarized in ►Table 5.
ing time was 134 minutes. The remaining data on surgery as
well as a comparison between the groups with and without
Discussion
anti-TNF-α are summarized in ►Table 3.
Seven patients experienced postoperative complications Our data shows that preoperative therapy with anti-TNF-α
(11%). Four of those (9.3%) were in the cohort treated with did not influence the risk of infectious and noninfectious
anti-TNF-α and 3 patients (13.6%) were in the non-anti-TNF-α postoperative complications, readmission rate, or ROR in this
Table 2 Preoperative medications for the entire cohort
Medication Overall; n ¼ 65 Anti-TNF-α; n ¼ 43 (66.15%) No anti-TNF-α; n ¼ 22 (33.85%) p-Value

n (%) n (%) n (%)
Corticosteroids 5 (7.7) 1 (2.3) 4 (18.2) 0.0413
Immunesuppressant 56 (86.1) 39 (90.7) 17 (77.3) 0.2532
Abbreviation: TNF-α, tumor necrosis factor α.

Note: Data are presented as numbers (%).

Table 3 Surgical variables for the entire cohort
Surgical variable Overall; n ¼ 65 Anti-TNF-α; n ¼ 43 No anti-TNF-α; n ¼ 22 p-Value

(66.15%) (33.85%)
n (%) n (%) n (%)
Open 61 (94) 40 (93) 21 (95.5) 1.0
Laparoscopic 4 (6) 3 (7) 1 (4.5) 1.0
Conversion 0 0 0 1.0
Restorative 61 (94) 41 (95) 20 (91) 0.599
End-to-end anastomosis 61 (100) 41 (100) 20 (100) 0.599
Duration of operation, min (median, range) 134 (60–300) 132 (60–260) 137 (90–300) 0.9061
Abbreviation: TNF-α, tumor necrosis factor α.

Note: Data are presented as numbers (%) unless otherwise noted.
Table 4 Multivariable analysis of associations between anti-

precious time period with respect to the fact that the
TNF-α use and postoperative complications
European Medicines Agency (EMA) has already approved
the anti-TNF-α therapy as a standard part of conservative
Variable OR 95% CI p-Value
therapy for pediatric CD patients.
Anti-TNF treatment 1.09 0.13–10.8 0.93 Anti-TNF-α therapy has become a standard part of man-
Age at procedure 0.81 0.54–1.14 0.22 aging severe CD forms in the pediatric patients, but the
Disease behavior 2.80 0.22–30.71 0.39 length of time it has been in use is still short, and its impact
on postoperative complications has not yet been analyzed
Acute complications 12.90 1.59–132.70 0.017
(infliximab, Remicade, and adalimumab, Humira, were first
Abbreviations: CI, confidence interval; GI, gastrointestinal; OR, odds approved by the EMA and the Food and Drug Administration
ratio; TNF-α, tumor necrosis factor α. [FDA] for pediatric CD in 2006 and 2014, respectively).
Notes: Disease behavior is according to the Paris classification – B3 versus
Guidelines from 2018 recommend anti-TNF-α therapy
B1 þ B2; Acute complications – GI bleeding, intestinal perforation, or
obstruction. should be discontinued prior to surgery in the adult CD
population, since its continued use is associated with higher
relatively large consecutive pediatric CD population. Our rates of postoperative septic complications.16 This finding
study includes the largest cohort of consecutive CD patients led ESPGHAN to discourage the use of anti-TNF-α during
treated with anti-TNF-α therapy for the shortest and most the perioperative period in pediatric patients.17 However,
Table 5 Postoperative complications and follow-up variables for the entire cohort
Variable Overall; n ¼ 65 Anti-TNF-α; n ¼ 43 No anti-TNF-α; n ¼ 22 p-Value

(66.15%) (33.85%)
n (%) n (%) n (%)
Postoperative length of stay, days (median, IQR) 11 (9–15) 10 (9–14) 11 (9–16) 0.0366
Mortality 0 0 0 1.0
SSI (S, D, OS, separation of stoma) 1(1.5) 1 (2.3) 0 1.0
Non-SSI (UTI, pneumonia, bacteremia) 0 0 0 1.0
D-C: minor complication 2 (3.1) 1 (2.3) 1 (4.6) 1.0
D-C: major complication 5 (7.7) 3 (7.0) 2 (9.1) 1.0
Ileus 2 (3.1) 1 (2.3) 1 (4.5) 1.0
Bleeding 1 (1.5) 1 (2.3) 0 (0) 1.0
Readmission 0 0 0 1.0
ROR 2 (3.1) 1 (2.3) 1 (4.5) 1.0
Follow-up, months (median, IQR) 23 (7 - 41) 26 (7 - 44) 20 (9 - 41) 0.41
IASC 3 (4.62) 1 (2.33) 2 (9.1) 0.26
Abbreviations: D, deep; D-C, the Clavien-Dindo classification (minor – I or II); IASC, intra-abdominal septic complication; Ileus, bowel obstruction with
radiographic confirmation and requirement of a nasogastric tube for decompression or operative intervention; IQR, interquartile range; non-SSI, non-
surgical site infections; OS, organ/space; ROR, return to the operating room; S, superficial; SSI, surgical site infection; UTI, urinary tract infection.
Note: Data are presented as numbers (%) unless otherwise noted.

previously published pediatric data did not show any evi- tious postoperative complications, readmission, or ROR in
dence supporting potential complications, and the trials the pediatric CD population. According to our study, there is
with adult and pediatric patients differed in several aspects. no need to discontinue anti-TNF-α during the preoperative
A single institution study published by Lightner et al evalu- period, but larger, prospective, and multicenter studies are
ated 30-day postoperative morbidity in a group of 69 patients; needed to validate this supposition.
54 patients received anti-TNF-α therapy within the 12 weeks
prior to surgery. The authors concluded that preoperative Conflict of Interest
exposure to anti-TNF-α does not influence the rate of overall None declared.
or infectious complications.18 A study published by Abbas et al
from Houston presented similar results. Postoperative mor-
bidity was similar in patients who were (N ¼ 22) or were not References
(N ¼ 51) treated with infliximab preoperatively (9% vs. 10%; 1 Gasparetto M, Guariso G. Highlights in IBD epidemiology and its
p ¼ 0.924).19 Zimmerman et al published a trial in 2016, which natural history in the paediatric age. Gastroenterol Res Pract
focused on postoperative complications in two patient groups, 2013;2013:829040
2 de Bie C, Paerregaard A, Kolacek S, et al. Disease phenotype at
based on whether their preoperative management included
diagnosis in pediatric Crohn&s disease. : 5-year analyses of the
infliximab (24 patients) or not (99 patients). There was no EUROKIDS Registry. Inflamm Bowel Dis 2013;19(02):378–385
statistically significant difference in postoperative morbidity 3 Mowat C, Cole A, Windsor A, et al; IBD Section of the British
between the two groups (21% vs. 11%; p ¼ 0.2).20 However, in Society of Gastroenterology. Guidelines for the management of
these last two studies, the standard use of anti-TNF-α therapy inflammatory bowel disease in adults. Gut 2011;60(05):571–607
4 Hanauer SB, Feagan BG, Lichtenstein GR, et al; ACCENT I Study
exceeded FDA/EMA permission, and both studies focused
Group. Maintenance infliximab for Crohn’s disease: the ACCENT I
solely on infliximab. The most recent work on this topic, which
randomised trial. Lancet 2002;359(9317):1541–1549
combined the entire inflammatory bowel disease (IBD) popu- 5 Hanauer SB, Sandborn WJ, Rutgeerts P, et al. Human anti-tumor
lation at the author’s hospital and evaluates postoperative necrosis factor monoclonal antibody (adalimumab) in Crohn’s dis-
complications in the entire biologics range, was published by ease: the CLASSIC-I trial. Gastroenterology 2006;130(02):323–333
authors from the United States and included 62 pediatric 6 Colombel JF, Sandborn WJ, Rutgeerts P, et al. Adalimumab for
maintenance of clinical response and remission in patients with
patients (40 with CD). Thirty-seven patients (18 CD patients)
Crohn’s disease: the CHARM trial. Gastroenterology 2007;132
received biologics preoperatively; however, this treatment did (01):52–65
not influence the rate of postoperative complications.21 Previ- 7 Rutgeerts P, Feagan BG, Lichtenstein GR, et al. Comparison of
ously, the association between anti-TNF-α and postoperative scheduled and episodic treatment strategies of infliximab in
complications was controversial,22,23 but recent studies in Crohn’s disease. Gastroenterology 2004;126(02):402–413
adults and subsequent meta-analysis demonstrate an increas- 8 Sandborn WJ, Rutgeerts P, Enns R, et al. Adalimumab induction
therapy for Crohn disease previously treated with infliximab: a
ing prevalence of postoperative complications, both infectious
randomized trial. Ann Intern Med 2007;146(12):829–838
and noninfectious.24 9 Vernier-Massouille G, Balde M, Salleron J, et al. Natural history of
A consensus on how to uniformly define and grade pediatric Crohn’s disease: a population-based cohort study.
different types of postoperative complications in pediatric Gastroenterology 2008;135(04):1106–1113
patients is still lacking, and therefore, there is currently no 10 Xu Y, Yang L, An P, Zhou B, Liu G. Meta-analysis: the influence of
background for further studies. We thus used the D-C system preoperative infliximab use on postoperative complications of
Crohn’s disease. Inflamm Bowel Dis 2019;25(02):261–269
in our analysis. This classification is a widely accepted tool
11 Chang M, Cohen B, Greenstein A. A Review of the Impact of
used to assess and report all postoperative complications, Biologics on Surgical Complications in Crohn’s Disease. Inflam-
but mainly in general surgery. None of the above-mentioned matory Bowel Diseases 2015;21(06):1472–1477
studies used this classification. Rather, they applied several 12 Rosenfeld G, Qian H, Bressler B. The risks of post-operative
heterogeneous systems. complications following pre-operative infliximab therapy for
Crohn’s disease in patients undergoing abdominal surgery: a
The main limitation of our study is its retrospective design.
systematic review and meta-analysis. J Crohn’s Colitis 2013;7
Furthermore, the study was conducted in the referral gastro- (11):868–877
enterological and surgical center, which may contribute to the 13 Levine A, Koletzko S, Turner D, et al. The ESPGHAN revised Porto
loss of diversity within the monitored group. criteria for the diagnosis of inflammatory bowel disease in
On the other side, pediatric IBD surgery should be per- children and adolescents. J Pediatr Gastroenterol Nutr 2014;58
formed preferably in high-volume centers where cooperation (06):795–806
14 Ruemmele FM, Veres G, Kolho KL, et al; European Crohn’s and
between gastroenterologists and surgeons is well established.
Colitis Organisation; European Society of Pediatric Gastroenter-
The data collection period is short, but this was intentional ology, Hepatology and Nutrition. Consensus guidelines of
to include the patients only within the time when anti-TNF-α ECCO/ESPGHAN on the medical management of pediatric Crohn’s
treatment was officially approved by the EMA for pediatric CD disease. J Crohn’s Colitis 2014;8(10):1179–1207
population. 15 Dindo D, Demartines N, Clavien PA. Classification of surgical
complications: a new proposal with evaluation in a cohort of
6336 patients and results of a survey. Ann Surg 2004;240(02):
Conclusion 205–213
16 Bemelman W, Warusavitarne J, Sampietro G, et al. ECCO-ESCP
Our data demonstrate that preoperative therapy with anti- consensus on surgery for Crohn’s disease. J Crohn’s Colitis 2018;
TNF-α did not increase the risk of infectious and noninfec- 12(01):1–16

17 Amil-Dias J, Kolacek S, Turner D, et al; IBD Working Group of 21 Mitsuya JB, Gonzalez R, Thomas R, El-Baba M. The effect of biologics
ESPGHAN (IBD Porto Group). Surgical management of Crohn on postoperative complications in children with inflammatory
disease in children: guidelines from the paediatric IBD Porto bowel disease and bowel resection. J Pediatr Gastroenterol Nutr
group of ESPGHAN. J Pediatr Gastroenterol Nutr 2017;64(05): 2019;68(03):334–338
818–835 22 Nasir BS, Dozois EJ, Cima RR, et al. Perioperative anti-tumor necrosis
18 Lightner AL, McKenna NP, Alsughayer A, et al. Anti-TNF biologic factor therapy does not increase the rate of early postoperative
therapy does not increase postoperative morbidity in pediatric complications in Crohn’s disease. J Gastrointest Surg 2010;14(12):
Crohn’s patients. J Pediatr Surg 2019:S0022-3468(19)30015-6 1859–1865
19 Abbas PI, Peterson ML, Fallon SC, et al. Evaluating the impact of 23 Syed A, Cross RK, Flasar MH. Anti-tumor necrosis factor therapy is
infliximab use on surgical outcomes in pediatric Crohn’s disease. associated with infections after abdominal surgery in Crohn’s
J Pediatr Surg 2016;51(05):786–789 disease patients. Am J Gastroenterol 2013;108(04):583–593
20 Zimmerman LA, Saites CG, Bairdain S, et al. Postoperative com- 24 Yang ZP, Hong L, Wu Q, Wu KC, Fan DM. Preoperative infliximab use
plications in children with Crohn disease treated with infliximab. and postoperative complications in Crohn’s disease: a systematic
J Pediatr Gastroenterol Nutr 2016;63(03):352–356 review and meta-analysis. Int J Surg 2014;12(03):224–230

Original Article
Investigation of Bosentan’s Effects on Pulmonary

Contusion Created by Blunt Thoracic Trauma in Rats
Gonca Gercel1 Burhan Aksu1 Seyma Ozkanli2 Hafize Uzun3 Feyza Aksu4 Erdem Ozatman1
Çiğdem Ulukaya Durakbaşa1
1 Department of Pediatric Surgery, Istanbul Medeniyet University Address for correspondence Çiğdem Ulukaya Durakbaşa, MD,
Göztepe Training and Research Hospital, Istanbul, Turkey Department of Pediatric Surgery, Medeniyet University Goztepe
2 Department of Pathology, Istanbul Medeniyet University Göztepe Training and Research Hospital, Istanbul 34000, Turkey
Training and Research Hospital, Istanbul, Turkey (e-mail: [email protected]).
3 Department of Biochemistry, Istanbul University Cerrahpasa Faculty
of Medicine, Istanbul, Turkey
4 Department of Cardiology, Istanbul Medeniyet University Göztepe
Training and Research Hospital, Istanbul, Turkey
Eur J Pediatr Surg
Abstract Introduction Bosentan is an endothelin-1 receptor antagonist with anti-inflammatory,

antioxidant, and antiproliferative effects. We aimed to evaluate its effects on lung tissue in a
pulmonary contusion (PC) model.
Materials and Methods The rats were randomly divided into five groups: PC3: PC
evaluated on the 3rd day (n ¼ 8), PC-B3: PC enteral bosentan 100 mg/kg/day, for 3 days
(n ¼ 8), PC7: PC evaluated on the 7th day (n ¼ 7), PC-B7: PC 7 days bosentan
100 mg/kg/day, for 7 days (n ¼ 8), C: control (n ¼ 6). Unilateral lung contusion was
created by dropping a metal weight onto the chest. The rats were sacrificed on the 3rd
or the 7th days. The lung tissue was evaluated histopathologically for alveolar edema,
congestion, and leukocyte infiltration, biochemically for malondialdehyde (MDA),
superoxide dismutase (SOD), and nitric oxide (NO) levels, and immunohistochemically
for inducible nitric oxide synthase (iNOS), endothelial nitric oxide synthase (eNOS), and
apoptosis scores.
Results Alveolar edema, congestion, and leukocyte infiltration scores were increased
in all groups compared with the control group (p < 0.05) and decreased in bosentan-
treated groups compared with the relevant nontreated groups (p < 0.05). Fibrosis was
observed only in PC7 and PC-B7 groups. Bosentan did not have any effect on fibrosis
development. iNOS and eNOS levels were higher in all groups compared with the
Keywords control (p < 0.05) without a difference in the nontreated versus treated groups
► pulmonary contusion (p > 0.05). Bosentan treatment caused decreased MDA and increased SOD levels in
► experimental comparison to the nontreated groups (p < 0.05). Tissue NO levels did not show any
► bosentan significant difference among groups. PC groups had higher levels of apoptosis
► rat compared with the control group (p < 0.05). The degree of apoptosis decreased in
► blunt thoracic trauma bosentan-treated groups compared with the nontreated groups (p < 0.05).

August 20, 2019
Investigation of Bosentan’s Effects on Pulmonary Contusion Gercel et al.
Conclusion PC causes progressive lung tissue damage. Bosentan reduced leukocyte

infiltration and alveolar edema and congestion caused by PC. It also decreased MDA
levels and increased SOD levels. Bosentan prevents tissue damage by inhibiting acute
inflammatory response and reduces oxidative stress secondary to inflammation. It has
therapeutic effects on apoptosis.
Introduction Animals and Experimental Protocol

Thirty-eight Sprague-Dawley rats, weighing 200 to 250 g,
Pulmonary contusion (PC) is a common type of injury that were included in this study under standard laboratory con-
occurs in 30 to 75% of thoracic injuries.1 The most common ditions (22 1°C, 12-hour light/dark cycle). They were fed
cause is blunt trauma. The pathophysiology of PC includes with standard rat chow and tap water ad libitum.
progressive inflammation, increased alveolar-capillary per- Rats were randomly assigned to one of the five groups: (1)
meability, pulmonary edema, ventilation/perfusion imbal- PC3: PC group (3 days) (n ¼ 8), (2) PC-B3: PC þ 3-day bosentan
ance, and loss of compliance.2 After the initial damage casued group (n ¼ 8), 100 mg/kg bosentan was given orally once a day
by the trauma itself, a secondary injury begins to develop in for 3 days, (3) PC7: PC group (n ¼ 7), (4) PC-B7: PC þ 7-day
the early period due to oxidative stress and inflammatory bosentan group (n ¼ 8), 100 mg/kg bosentan was given
response.3 PC results in programmed cell death and an orally once a day for 7 days, and (5) C: control group (n ¼ 6).
ineffective epithelial repair process can result in tissue The rats were anesthetized with intraperitoneal keta-
fibrosis.4,5 Therefore, early and effective treatment should mine (50 mg/kg) and xylazine (15 mg/kg). PC was induced
be initiated for managing patients with PC. by dropping a cylindrical metal weight (0.4 kg) from a
Endothelin (ET) is derived from vascular smooth muscle specified distance (60 cm) through a stainless steel tube
endothelium. It is the most potent vasoconstrictor molecule onto the right hemithorax. In this way, the trauma was
contributing to inflammation, smooth muscle proliferation, standardized by applying 2.35 J energy on the chest accord-
and fibrosis formation.6 ET-1 is the most abundant isoform of ing to the formula E ¼ mgh (E ¼ energy [joule], m ¼ mass of
ET and produced by smooth muscle cells, macrophages, the cylinder [kg], g ¼ gravity constant [9.8 m/s2], and
airway epithelium, and alveolar epithelial cells. ET-1 causes h ¼ height [meter]). This is a modified form of the model
fibrosis of vascular cells, induces proinflammatory mecha- defined by Raghavendran et al.13
nisms, and stimulates production of reactive oxygen species A biopsy of the right lung tissue was performed thorough
and cytokine secretion.7 a thoracotomy incision on either the 3th or the 7th day of the
Bosentan is an orally active, highly substituted pyrimidine experiment depending on the allocated groups. Tissue sam-
derivative. It acts as a mixed and competitive antagonist of ples were fixed in 10% formaldehyde for histopathologic
ET-1 at the ET-A and ET-B receptors. It shows anti-inflam- examination and frozen at –80°C for subsequent biochemical
matory, antioxidant, and antiproliferative effects by blocking assay. The rats were humanely killed by bleeding from the
these receptors.8 Bosentan is in clinical use for various thoracic aorta.
diseases in humans. It is used in the treatment of pulmonary
arterial hypertension (PAH), PAH associated with congenital Histopathological Examination
heart disease, and for prevention of digital ulcers in systemic Paraffin-embedded tissues were stained with routine hema-
sclerosis, regardless of the presence of PAH.9 In addition, toxylin and eosin and Masson-Trichrome stain to be exam-
possible beneficial effects of bosentan have been explored in ined under a light microscope (Olympus BX-51) by the same
various disorders, such as chronic heart failure, essential pathologist blinded to the study. The biopsies were examined
hypertension, subarachnoid hemorrhage, cerebral vaso- for alveolar edema, congestion, leukocyte infiltration, and
spasm, migraine, and prevention of damage to gastric muco- fibrosis.
sa by acetylsalicylic acid.10 Alveolar edema and congestion levels were determined by
The efficacy of bosentan was evaluated previously in exper- scoring from 0 to 3.14 Grade 0: no pathology (0–5%), Grade 1:
imental aspiration pneumonia and lung emphysema models mild (6–15%), Grade 2: moderate (16–20%), and Grade 3:
and it was found to be effective in suppressing the inflamma- severe (21–25%).
tory response in the airways.11,12 There is no study evaluating Leukocyte infiltration was graded as follows14: Grade 0:
the effects of bosentan in PC. Therefore, the present study no extravascular leukocytes, Grade 1: leukocyte count < 10,
aimed to evaluate the effects of bosentan on lung injury caused Grade 2: leukocyte count between 10 and 45, and Grade 3:
by blunt thoracic trauma with resultant PC. leukocyte count > 45.
Fibrosis was scored from 0 to 4.15 Score 0: normal lung
architecture, Score 1: increased thickness of some ( 50%) of
interalveolar septa, Score 2: thickening of > 50% of interalveolar
The experiments were performed in adherence to the approval septa without formation of fibrotic foci, Score 3: thickening of
of the Ethical Committee of Yeditepe University (2017/601). the interalveolar septa with formation of isolated fibrotic foci,

and Score 4: formation of multiple fibrotic foci with total or at 3,000 g at 40°C to remove the debris. Clear upper super-
subtotal distortion of parenchymal architecture. natant was taken and the determination of superoxide dis-
mutase (SOD), malondialdehyde (MDA), and nitric oxide (NO)
Immunohistochemical Analysis was performed.
Terminal deoxynucleotidyl transferase dUTP nick end labeling Superoxide dismutase activity was determined by the
(TUNEL), inducible nitric oxide synthase (iNOS), and endothelial method modified by Sun et al.18 Cytosolic superoxide dis-
nitric oxide synthase (eNOS) were studied immunohistochemi- mutase (Cu/Zn SOD) containing copper (Cu) and zinc (Zn)
cally and the evaluation was performed by a single pathologist was measured. MDA, the final product of lipid peroxidation,
under a light microscope (Olympus BX-51). was measured using the method modified by Ohkawa et al.19
Apoptotic cells were stained by TUNEL technique using an The concentration of thiobarbituric acid-reactive agents was
apoptosis detection kit (ApopTag Plus Peroxidase In Situ calculated using the 1.56 10–5 M1 cm1 mol/L extinction
Apoptosis Kit, Code: S7101, EMD Millipore, Temecula, Cal- coefficient. Tissue NO levels were measured by a commercial
ifornia, United States). In the evaluation of TUNEL staining, spectrophotometric kit (BioVision, Incorporated, 155 S. Mil-
nuclear staining was accepted as positive. pitas Boulevard, Milpitas, California, United States).
iNOS antibody kit (Primary Antibody: iNOS, Clone: Poly-
clonal, Code: E3740, SPRING Bioscience, Pleasanton, California, Statistical Analyses
United States, secondary kit: LEICA-UK, Bond Polymer Refine Number Cruncher Statistical System 2007 (Kaysville, Utah,
Detection wDAB, IHC Stainer: LEICA-UK, Bond Max IHC Stainer) United States) program was used for statistical analysis.
was used to determine iNOS activity immunohistochemically. Descriptive statistical methods (mean, standard deviation,
In the evaluation of iNOS staining, cytoplasmic staining was median, frequency, ratio, minimum, maximum) were used
accepted as positive. when evaluating the study data. Kruskal–Wallis test was
The number of iNOS and TUNEL-positive cells in each used for comparison of three or more groups which did not
specimen was also scored.16 Five to 10 randomly selected show normal distribution and Bonferroni-corrected Mann–
areas were scored for each specimen in every experiment as Whitney U test was used to determine the group that caused
follows: 0 ¼ no positive response, 1 ¼ 10% of cells, 2 ¼ 11 to the difference with p-values less than 0.05 considered to be
20% of cells, 3 ¼ 21 to 40% of cells, and 4 ¼ > 40% of cells. significant.
These analyses were performed in two sections from each
animal at 40 magnification in at least 10 different regions
Results
for each section.
A semiquantitative scale was used for quantification of Histopathological Findings
immunohistochemical expression of eNOS protein in lung There was a statistically significant difference in alveolar edema
tissue by eNOS antibody kit (Primary Antibody: eNOS, Clone: and congestion, and leukocyte infiltration scores of the groups
Polyclonal, Code: E4110, SPRING Bioscience, secondary kit: (p < 0.01) (►Table 1). The scores of PC3, PC7, PC-B3, and PC-B7
LEICA-UK, Bond Polymer Refining Detection wDAB, IHC groups were significantly higher than those of the control
Stainer: LEICA-UK, Bond Max IHC Stainerr): Grade 0 ¼ no group (p < 0.05). The relevant scores of the two treatment
staining, Grade 1 ¼ slightly staining, Grade 2 ¼ moderately groups, PC-B3 and PC-B7, were significantly lower than the
staining, and Grade 3 ¼ intensely staining.17 In the evalua- nontreated respective PC3 and PC7 groups (p < 0.05) (►Fig. 1).
tion of eNOS staining, cytoplasmic staining was accepted as The fibrosis scores of the groups significantly differed
positive. among groups (p < 0.01). Fibrosis development was not
observed on the 3rd day either in the PC3 or in PC-B3 groups.
Biochemical Examination It was present both in the PC7 and PC-B7 groups. There was
The lung tissues were homogenized using 20% phosphate no significant difference in the fibrosis scores between the
buffer via the homogenizer (Next Advance Bullet Blender PC-B7 group treated with bosentan and the nontreated PC7
Storm 24). The homogenates were centrifuged for 10 minutes group (p > 0.05) (►Fig. 2).
Table 1 Alveolar edema/congestion, leukocyte infiltration, and fibrosis scores
Rats PC3 (n ¼ 8) PC7 (n ¼ 7) PC-B3 (n ¼ 8) PC-B7 (n ¼ 8) Control (n ¼ 6)

Alveolar edema/congestion 2.50 0.53a 2.00 0.58a 1.50 0,76a,c 1.25 0.46a,d 00
a b b,c b,d
Leukocyte infiltration 2.88 0.35 2.29 0,76 1.63 0.92 1.38 0.74 0.33 0.52
b,c b,e
Fibrosis score 00 0.86 0.90 00 0.63 0.74 00
Abbreviations: B, bosentan; PC, pulmonary contusion.

Note: Values are expressed as mean standard deviation Mann–Whitney U-test.
a
p < 0.01 with respect to control group.
b
c
p < 0.05 with respect to PC3 group.
d
e
p > 0.05 with respect to PC7 group.

Fig. 1 (A) Control group showing normal lung tissue without contusion [alveolar space (fine arrow), alveolar septa (arrow head), bronchiole
(thick arrow), and vascular structure (star)]. (B) PC3 group showing intense alveolar edema and congestion and leukocyte infiltration (arrows)
(score 4) [bronchiole structure ()]. (C) PC-B3 group showing less concentrated alveolar edema and congestion, and leukocytic infiltration
(score 1). (D) PC7 group showing alveolar edema and congestion and leukocytic infiltration (score 2). (E) PC-B7 group showing less concentrated
alveolar edema and congestion and leukocytic infiltration (score 1) (hematoxylin and eosin [H&E]) (10).
Fig. 2 (A) Control group showing normal lung tissue without contusion. There is no staining with Masson-Trichrome stain (MTC) [alveolar septa
(thin arrows), bronchiole (thick arrow), and vessel structure (star)] (MTC) (10). (B) PC7 group showing lung tissue sampled from rats 7 days
after the contusion [MTC-stained intra-alveolar septa thickening and collagen deposition (arrows)] (MTC) (40).

Table 2 Comparison of superoxide dismutase, malondialdehyde, and nitric oxide values

a a b d
SOD 6.86 0.87 6.54 0.49 8.12 1.27 7.22 0.49 7.74 0.59
MDA 2.29 0.53a 10.02 2.00a,c 2.05 0.55e 3.03 0.80d 1.86 0.64
NO 64.69 18.42f 71.19 9.27f 61.51 15.89f 65.51 14.40f 62.4 13.17
Abbreviations: B, bosentan; MDA, malondialdehyde; NO, nitric oxide; PC, pulmonary contusion; SOD, superoxide dismutase.
Note: Values are expressed as mean standard deviation Mann–Whitney U-test.
a
b
c
d
e
f
p > 0.05 with respect to control group.
Biochemical Findings group (p > 0.05). eNOS scores of all four PC groups were
There was a statistically significant difference in SOD and found to be significantly higher than the control group
MDA levels among groups (p < 0.01) (►Table 2). The non- (p < 0.05). The comparison of eNOS scores between PC3
treated PC3 and PC7 groups had lower SOD levels than the and PC-B3 or PC7 and PC-B7 groups did not reveal statistically
control group (p < 0.05). The SOD levels of the bosentan- significant difference (p > 0.05) (►Fig. 3).
treated PC-B3 and PC-B7 groups did not show statistically In the evaluation of apoptosis by TUNEL staining, a statisti-
significant difference when compared with the control group cally significant difference was found among groups (p < 0.01).
(p > 0.05). The SOD levels obtained in the PC-B3 group were TUNEL scores of the PC3 and PC7 groups were significantly
higher than in PC3 (p < 0.05). Similarly, the levels obtained in higher than those of the control group (p < 0.05). Comparison
PC-B7 were higher than in PC7 4 (p < 0.05). of TUNEL scores between the PC3 and PC-B3 and PC7 and PC-B7
MDA levels were higher in the PC3 and PC7 groups in groups revealed significant difference, with the bosentan-
comparison to the control group (p < 0.05). The PC7 group treated groups having lower scores (p < 0.05) (►Fig. 4).
had higher MDA levels when compared with the PC3 group
(p < 0.01). When the nontreated groups are compared
Discussion
against the respective bosentan-treated groups, MDA levels
did not show statistically significant difference between the Blunt chest trauma is commonly associated with PC leading to
PC3 and PC-B3 groups (p > 0.05) but they were decreased in high morbidity and mortality.1,2 Prevention of progressive
the PC-B7 group in comparison to PC7 (p < 0.05). pulmonary damage caused by PC has a critical role in the
Tissue NO levels did not show any statistically significant prognosis of patients, especially when performed as early as
difference among groups (p > 0.05). possible.20 In daily clinical practice, the treatment alternatives
for PC are oxygen support, close observation, and management
Immunohistochemical and TUNEL Findings of complications, because there is no standard and disease-
iNOS and eNOS scores showed statistically significant differ- specific treatment.21 Developing new strategies or agents to
ence among groups (p < 0.05) (►Table 3). The PC3 group had prevent progression and complications of PC is of vital
higher iNOS scores than both the PC-B3 and control groups importance.
(p < 0.05). The relevant scores were not significantly differ- ET is a proinflammatory vasoactive peptide that induces
ent between the PC7 group and the bosentan-treated PC-B7 the production of cytokines and reactive oxygen products
Table 3 iNOS, eNOS, and TUNEL scores

iNOS score 2.25 1.49a 1.43 1.27a 0.88 0.64c 1.50 0.93f 0.33 0.82
a a b,e a,f
eNOS score 2.00 0.93 1.86 0.69 1.75 0.46 1.50 0.53 0.17 0.41
a a c d
TUNEL score 2.63 1.51 2.14 1.86 1.00 1.31 0.13 0.35 0.17 0.41
Abbreviations: B, bosentan; eNOS, endothelial nitric oxide synthase; iNOS, inducible nitric oxide synthase; PC, pulmonary contusion; TUNEL, terminal
deoxynucleotidyl transferase dUTP nick end labeling..
Note: Values are expressed as mean standard deviation Mann–Whitney U test.
a
b
c
d
e
f

Fig. 3 (A) Control group showing inducible nitric oxide synthase (iNOS) staining in normal lung tissue (iNOS) (20). (B) PC3 group showing
findings 3 days after the contusion with brown cytoplasmic iNOS staining (arrows) (iNOS) (20). (C) Control group showing eNOS staining in
normal lung tissue (endothelial nitric oxide synthase [eNOS]) (40). (D) PC3 group showing findings 3 days after the contusion with brown
cytoplasmic eNOS staining (arrows) (eNOS) (40).
from polymorphonuclear neutrophils which can be activated on day 3 with the maximum levels being reached on day 21 in
in various types of lung injury.22 On the other hand, bosentan the injury group. In the bosentan-treated group, a significant
is an agent showing anti-inflammatory, antioxidant, and decrease in HYP levels was found on the 21st day. In the present
antiproliferative effects by blocking ET receptors.8 The pres- study, although there was no evidence of fibrosis in any group
ent study evaluated anti-inflammatory and antioxidant at day 3, it was evident by Masson-Trichrome stain on day 7.
effects of bosentan as well as its effects on fibrosis in a PC However, bosentan treatment did not alter fibrosis scores. The
model induced by blunt thorax trauma. absence of fibrosis on the 3rd day in our study may be related to
It was previously shown that parenchymal and perivascular the technique of evaluation of fibrosis. Additionally, bosentan
neutrophil infiltration and atelectasis with alveolar surface was ineffective in improving fibrosis probably because a 7-day
changes start 24 hours after trauma and increase significantly treatment course was not enough. On a longer term, the results
after 48 hours.23 The lung parenchymal injury and the bron- regarding bosentan’s effects on fibrosis could be different.
chial damage occurs at the end of 48 hours with the fibroblas- It is known that alveolar macrophages can produce potent
tic activity occurring in the late period. Inflammation increases superoxide radicals in cases of acute lung injury including PC.26
over days and the damage process continues after trauma. The Bosentan has previously been shown to reduce oxidative
present study showed signs of tissue disruption such as damage as indicated by increased MDA and decreased SOD
alveolar edema, congestion, and leukocyte infiltration in all levels in different pathologies.27,28 The present study sup-
trauma groups. These effects were similar on the 3rd and 7th ported literature findings with high MDA and low SOD levels in
days and inflammation persisted on the 7th day with the same the trauma groups which were significantly improved by
intensity as in the 3rd day. These findings support that the bosentan treatment.
inflammatory findings in the traumatic PC continue for a long NO plays an important role in inflammation and oxidative
time after trauma. Leucocyte infiltration and alveolar edema reactions resulting from trauma-induced PC.29 Agents that
and congestion were significantly lower in the groups treated modulate NO activity are considered to have a therapeutic
with bosentan proving its anti-inflammatory effects in the value.30,31 All three isoforms, eNOS, neuronal nitric oxide
present PC model. synthase, and iNOS are present in the lungs.32 iNOS is known
An ineffective epithelial repair process with epithelial to be secreted excessively in the airway epithelium with the
damage may result in fibrosis in the tissues.5,24,25 Interstitial induction of specific cytokines after trauma but the studies on
inflammation and fibrosis start to increase on the 3rd day of eNOS response after trauma are contradictory.33,34 Unlike
acute lung injury and increase with a faster acceleration after chemical lung injury models in the literature, the present
the 7th day.24 In an acute lung injury model by using the study found that eNOS levels were significantly higher in the
herbicide paraquat, pulmonary fibrosis development was contusion group compared with the control group.33 It is likely
evaluated with hydroxyproline (HYP) levels on the 3rd, 7th, that direct injury and disruption of endothelial integrity in
14th, and 21st days and also the efficacy of bosentan on blunt thoracic trauma results in an increase in eNOS levels. On
pulmonary fibrosis was assessed.25 HYP levels were increased the other hand, in accordance with the literature, the levels of

Fig. 4 (A) Control group showing terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) staining in normal lung tissue (score 0)
(TUNEL) (20). (B) PC3 group showing apoptotic staining with TUNEL 3 days after the contusion with > 40% TUNEL-positive apoptotic cells
(arrows) (score 4) (TUNEL) (40). (C) Bosentan-treated PC-B3 group showing 10% TUNEL-positive cells (arrow) (score 1) (TUNEL) (40). (D) PC7
group showing > 40% TUNEL-positive apoptotic cells (arrows) 7 days after the contusion (score 4) (TUNEL) (20). (E) PC-B7 group showing no
positive TUNEL staining (score 0) (TUNEL) (20).
iNOS in lung tissue were significantly higher in the contusion TUNEL positive cells were significantly higher in the contusion
group. The effects of ET receptor antagonists including bosen- groups than in the control group confirming that PC results in
tan on NO pathway is unclear.22,35 In the present study, apoptosis. Moreover, TUNEL scores were lower in the treat-
although both iNOS and eNOS scores were lower in the groups ment groups both on the 3rd and the 7th days than in the
receiving bosentan no statistically significant difference was contusion-only groups. Bosentan was effective in reducing
detected. Furthermore, tissue NO levels were not different apoptosis caused by PC in this model.
among the groups. Thus, bosentan neither suppressed the There are several limitations of the current study. First, the
synthesis of iNOS or eNOS nor had a significant effect on NO serum concentrations of NO, MDA, and SOD were not mea-
levels in our model. Considering its positive effects on MDA and sured. Second, the contralateral lungs were not evaluated.
SOD levels, bosentan seems to act on the oxidative process by Finally, the use of rat models does not adequately imitate
anti-inflammatory mechanisms. the clinical scenarios of human diseases. Yet, bosentan is an
Eukaryotic cells die either by way of pathological cell death Food and Drug Administration-approved drug that can be used
(necrosis) or programmed cell death (apoptosis).36 Apoptotic in different pathologies in humans. There is no study on the use
cell death is induced by oxidative stress, NO, and ischemia.37,38 of bosentan in PC after blunt trauma in the medical literature. It
Liener et al demonstrated the first evidence of the initiation of has a therapeutic potential to be used for prevention of PC-
programmed cell death in lung tissue following PC in an associated lung damage.
experimental study.39 They showed bilateral PC in trauma In conclusion, PC causes progressive lung tissue damage.
animals with higher numbers of apoptotic cells in lung but not Bosentan reduced leukocyte infiltration and alveolar edema
in liver tissue as early as 6 hours after the injury. They and congestion. It also decreased MDA levels and increased SOD
concluded that local mechanisms may be responsible for the levels. Bosentan prevents tissue damage by inhibiting acute
induction of programmed cell death. Apoptosis was assessed inflammatory response and reduces oxidative stress secondary
by TUNEL staining in the present study. The scores obtained by to inflammation. It has therapeutic effects on apoptosis.

Conflict of Interest 19 Ohkawa H, Ohishi N, Yagi K. Assay for lipid peroxides in animal
None declared. tissues by thiobarbituric acid reaction. Anal Biochem 1979;95
(02):351–358
20 Türüt H, Ciralik H, Kilinc M, Ozbag D, Imrek SS. Effects of early
References administration of dexamethasone, N-acetylcysteine and aproti-
1 Miller PR, Croce MA, Bee TK, et al. ARDS after pulmonary contusion: nin on inflammatory and oxidant-antioxidant status after lung
accurate measurement of contusion volume identifies high-risk contusion in rats. Injury 2009;40(05):521–527
patients. J Trauma 2001;51(02):223–228 21 Keough V, Pudelek B. Blunt chest trauma: review of selected
2 Cohn SM, Zieg PM. Experimental pulmonary contusion: review of pulmonary injuries focusing on pulmonary contusion. AACN Clin
the literature and description of a new porcine model. J Trauma Issues 2001;12(02):270–281
1996;41(03):565–571 22 Fujitani Y, Trifilieff A, Tsuyuki S, Coyle AJ, Bertrand C. Endothelin
3 Lang JD, McArdle PJ, O’Reilly PJ, Matalon S. Oxidant-antioxidant receptor antagonists inhibit antigen-induced lung inflammation
balance in acute lung injury. Chest 2002;122(6, Suppl):314S–320S in mice. Am J Respir Crit Care Med 1997;155(06):1890–1894
4 Seitz DH, Perl M, Mangold S, et al. Pulmonary contusion induces 23 Raghavendran K, Notter RH, Davidson BA, Helinski JD, Kunkel SL,
alveolar type 2 epithelial cell apoptosis: role of alveolar macro- Knight PR. Lung contusion: inflammatory mechanisms and inter-
phages and neutrophils. Shock 2008;30(05):537–544 action with other injuries. Shock 2009;32(02):122–130
5 Raghavendran K, Davidson BA, Woytash JA, et al. The evolution of 24 Bhargava M, Wendt CH. Biomarkers in acute lung injury. Transl
isolated bilateral lung contusion from blunt chest trauma in rats: Res 2012;159(04):205–217
cellular and cytokine responses. Shock 2005;24(02):132–138 25 Zhang Z, Jian X, Zhang W, Wang J, Zhou Q. Using bosentan to treat
6 Yanagisawa M, Kurihara H, Kimura S, et al. A novel potent paraquat poisoning-induced acute lung injury in rats. PLoS One
vasoconstrictor peptide produced by vascular endothelial cells. 2013;8:1–12
Nature 1988;332(6163):411–415 26 Gris D, Hamilton EF, Weaver LC. The systemic inflammatory
7 Bonifacea S, Reynaud-Gaubertb M. Endothelin receptor antago- response after spinal cord injury damages lungs and kidneys.
nists-their role in pulmonary medicine. Rev Mal Respir 2011; Exp Neurol 2008;211(01):259–270
28:94–107 27 Cayir A, Ugan RA, Albayrak A, et al. The lung endothelin system: a
8 Clozel M, Breu V, Gray GA, et al. Pharmacological characterization of potent therapeutic target with bosentan for the amelioration of
bosentan, a new potent orally active nonpeptide endothelin recep- lung alterations in a rat model of diabetes mellitus. J Endocrinol
tor antagonist. J Pharmacol Exp Ther 1994;270(01):228–235 Invest 2015;38(09):987–998
9 Rubin LJ, Badesch DB, Barst RJ, et al. Bosentan therapy for 28 Yayla M, Halici Z, Unal B, Bayir Y, Akpinar E, Gocer F. Protective effect
pulmonary arterial hypertension. N Engl J Med 2002;346(12): of ET-1 receptor antagonist bosentan on paracetamol induced acute
896–903 liver toxicity in rats. Eur J Pharmacol 2014;726:87–95
10 Dingemanse J, van Giersbergen PL. Clinical pharmacology of 29 Özkan M, Yüksekol İ. Nitrik Oksit ve Akciğerler. Toraks Derg 2003;
bosentan, a dual endothelin receptor antagonist. Clin Pharmaco- 4:88–94
kinet 2004;43(15):1089–1115 30 Lowenstein CJ, Dinerman JL, Snyder SH. Nitric oxide: a physiologic
11 Trabold B, Pawlik M, Nietsch R, et al. Bosentan reduces oxidative messenger. Ann Intern Med 1994;120(03):227–237
burst in acid aspiration-induced lung injury in rats. Injury 2009; 31 Kubes P, McCafferty DM. Nitric oxide and intestinal inflammation.
40(09):946–949 Am J Med 2000;109(02):150–158
12 Gamze K, Mehmet HM, Deveci F, Turgut T, Ilhan F, Ozercan I. Effect 32 Knowles RG, Moncada S. Nitric oxide synthases in mammals.
of bosentan on the production of proinflammatory cytokines in a Biochem J 1994;298(Pt 2):249–258
rat model of emphysema. Exp Mol Med 2007;39(05):614–620 33 Cox RA, Jacob S, Oliveras G, et al. Pulmonary expression of nitric
13 Raghavendran K, Davidson BA, Helinski JD, et al. A rat model for oxide synthase isoforms in sheep with smoke inhalation and burn
isolated bilateral lung contusion from blunt chest trauma. Anesth injury. Exp Lung Res 2009;35(02):104–118
Analg 2005;101(05):1482–1489 34 Minchenko AG, Armstead VE, Opentanova IL, Lefer AM. Endothe-
14 Karaca S. The effects of s-methylisothiourea treatment in rats lin-1, endothelin receptors and ecNOS gene transcription in vital
with unilateral lung contusion. Medical Thesis, Department of organs during traumatic shock in rats. Endothelium 1999;6(04):
Pediatric Surgery, Trakya University Faculty of Medicine, Edirne, 303–314
Turkey; 2013 35 Finsnes F, Skjønsberg OH, Tønnessen T, Naess O, Lyberg T, Chris-
15 Lawson WE, Polosukhin VV, Stathopoulos GT, et al. Increased and tensen G. Endothelin production and effects of endothelin antag-
prolonged pulmonary fibrosis in surfactant protein C-deficient onism during experimental airway inflammation. Am J Respir Crit
mice following intratracheal bleomycin. Am J Pathol 2005;167 Care Med 1997;155(04):1404–1412
(05):1267–1277 36 Kannan K, Jain SK. Oxidative stress and apoptosis. Pathophysiology
16 Ayvaz S, Aksu B, Karaca T, et al. Effects of methylene blue in acute 2000;7(03):153–163
lung injury induced by blunt chest trauma. Hippokratia 2014;18 37 Ikeda H, Suzuki Y, Suzuki M, et al. Apoptosis is a major mode of cell
(01):50–56 death caused by ischaemia and ischaemia/reperfusion injury to
17 Kaminski A, Pohl CB, Sponholz C, et al. Up-regulation of endothe- the rat intestinal epithelium. Gut 1998;42(04):530–537
lial nitric oxide synthase inhibits pulmonary leukocyte migration 38 Lee YB, Yune TY, Baik SY, et al. Role of tumor necrosis factor-alpha
following lung ischemia-reperfusion in mice. Am J Pathol 2004; in neuronal and glial apoptosis after spinal cord injury. Exp Neurol
164(06):2241–2249 2000;166(01):190–195
18 Sun Y, Oberley LW, Li Y. A simple method for clinical assay of 39 Liener UC, Knöferl MW, Sträter J, et al. Induction of apoptosis
superoxide dismutase. Clin Chem 1988;34(03):497–500 following blunt chest trauma. Shock 2003;20(06):511–516

Original Article
Surgical Management of Esophageal Achalasia

in Pediatrics: A Systematic Review
Ayman Goneidy1 James Cory-Wright2 Limeng Zhu2 Georgina Malakounides3
1 Department of Paediatric Surgery, Royal Manchester Children’s Address for correspondence Ayman Goneidy, MBBCh, MSc, MRCS,
Hospital, Manchester, United Kingdom Department of Paediatric Surgery, Royal Manchester Children’s
2 School of Medicine, University of Cambridge School of Clinical Hospital, Oxford Road, Manchester M13 9WL, United Kingdom
Medicine, Cambridge, United Kingdom (e-mail: [email protected]).
3 Department of Paediatric Surgery, Cambridge University Hospitals
NHS Foundation Trust, Cambridge, United Kingdom
Eur J Pediatr Surg
Abstract Introduction There are no evidence-based guidelines on the surgical management of

esophageal achalasia (OA) in children. This can be a challenging condition with
significant physical and psychological morbidity. Our aim was to identify the most
common management modalities and their outcomes.
Materials and Methods A systematic review was performed through a literature
search of health care databases in accordance with the Preferred Reporting Items for
Systematic Reviews and Meta-Analyses guidelines, aiming at identifying pediatric
series discussing the diagnosis and management of OA. Duplicates, case series with < 9
patients, and follow-up of < 1 year were excluded. The included papers were analyzed
for diagnostic methods, primary treatment method, complications, follow-up dura-
tion, outcome measures recorded, and outcome.
Results Data from 33 papers for 742 children treated for OA was analyzed. Eleven
mentioned multiple management modalities. In summary, 25 described Heller’s esoph-
agomyotomy (HM), 13 esophageal dilatation (EOD), and 6 peroral esophageal myotomy
(POEM). Mean follow-up was 43.7 months (12–180). Outcome measures were heteroge-
neous. However, analysis of reported success showed a mean success of 78% for HM
(p ¼ 1.79 10–7), 44.9% for EOD (p ¼ 0.24), and 99.3% for POEM (p ¼ 0.001). Reported
complications were 12.8% for HM, 5% for EOD, and 24.4% for POEM. Further interventions
Keywords were required for 10.9% of HM, 62.3% of EOD, and 0.01% of POEM patient groups.
► achalasia Conclusion Methods of diagnosis and measures of successful outcomes were heteroge-
► Heller neous, limiting the strength of evidence. HM showed superior short-term success rates to
► children EOD. POEM is a promising modality but requires investment in equipment and training.
► management Information about sustainability of response and long-term outcomes is lacking.
Introduction
year in the United Kingdom.3,4 Patients usually present with
Esophageal achalasia (OA) is a disease that affects the lower progressive dysphagia to fluids or solids, vomiting, regurgita-
esophageal sphincter (LES) that leads to dysfunctional esoph- tion, chest pain, or repeated chest infections.
ageal motility.1,2 It is an extremely rare disease in the pediatric Multiple investigations have been described for establishing
population with an incidence rate of 1.8:100,000 children per the diagnosis of OA. Contrast esophagogram is used to exclude a

August 20, 2019
Surgical Management of Esophageal Achalasia in Pediatrics Goneidy et al.
mechanical obstruction and show the delay in transition than 1 year, undefined outcome measures, non-English
through the LES (classical parrot beak appearance). Esophago- articles, theoretical or animal model studies, duplicated
scopy is used to rule out pseudo-achalasia and to assess the articles, and abstracts. Included papers were analyzed for
esophageal wall for inflammation. Manometry is also used to methods used for diagnosis, primary treatment method, fol-
assess the LES pressure and lower esophageal motility.5 low-up duration, outcome measures recorded, subsequent
Management of OA is aimed at providing symptomatic treatments, and complications. Patients were grouped into
relief. Treatments reportedly used in children include nifedi- three main categories according to their main treatment
pine,6 esophageal injection with botulinum toxin,7 Heller method: (1) HM group (open or minimally invasive), (2)
myotomy (HM) with or without an antireflux procedure, EOD group, and (3) POEM group.
endoscopic esophageal dilatation (EOD), and more recently
peroral esophageal myotomy (POEM).8,9 None of the men-
Results
tioned treatments are curative, and there is still no national or
international consensus on one gold standard intervention in A total of 33 complete studies were included for analysis after
children which provides sustainable benefit with minimal risk. applying filters of inclusion and exclusion criteria to the
Authors of this study undertook a systematic review to search studies resulting from the search (►Fig. 1). An illustration of
the literature for publications that describe diagnosis and the studies’ outlines is depicted in ►Table 1. Included studies
treatment of OA to compare outcomes and complication rates. were published between 1980 and 2018. None of the studies
were randomized controlled trials; 3 studies (9%) were
prospective and 11 studies (33%) compared outcomes of
two or more treatments. Almost all studies were case series
A systemic review was performed in accordance with the (Level 4 evidence).
Preferred Reporting Items for Systematic Reviews and Meta- Data for 742 children treated for OA was analyzed. Median
Analyses guidelines by two independent researchers. An number of patients was n ¼ 20 (9–86). Patients’ ages ranged
electronic search was done on the Medline (Pubmed) and between 3 months to 17 years. Fifty-five percent were male
EM Base (OVID) for all English publications about diagnosis patients. Presenting symptoms were dysphagia, regurgitation/
and management of OA in pediatrics. Search terms were vomiting, intolerance to solids or liquids, aspiration pneumo-
achalasia AND (children OR child OR childhood OR Pediatric nia, odynophagia, and/or loss of weight. Methods of diagnosis
OR pediatric) AND (treat OR treatment) OR (Manage OR were not mentioned in two studies. A contrast swallow and
management). Qualitative strength of studies was assessed manometry diagnostic endoscopy were the predominant
against critical appraisal hierarchy of evidence.10,11 investigations used for diagnosis in 28 studies (84.8%).
Exclusion criteria were studies with patients older than Primary method of treatment in 25 studies was HM, 13
18 years of age, achalasia not being the primary pathology, less described EOD, and 6 described POEM. Few studies described
than 9 or an unknown number of patients, follow-up of less other treatment modalities like botulinum injection (6%) and
Fig. 1 Flowchart illustrating the selection process for studies discussing esophageal atresia in children.

Table 1 Characteristics of studies included in the systematic review
Author Year N Age Investigations Main Follow-up Outcome measure

(range) treatment (mean/
average
years)
Azizkhan et al13 1980 19 0.5–15 Contrast-manometry HM-EOD 4 Symptoms-contrast-weight
Boyle et al14 1981 10 10–17 Contrast-manometry-scope EOD 1.5 Symptoms-weight
Vane et al15 1988 21 0.5–17 Contrast manometry HM 6.3 Symptoms
Nihoul-Fékété 1989 35 2 wk–15 Contrast- scope HM NS Symptoms-contrast-weight
et al16
Emblem et al17 1993 14 0.3–0.8 Contrast- scope HM-EOD- 3.9 Symptoms-contrast-
Nifedipine weight-scope
Lelli et al18 1997 19 1–17 Contrast-manometry HM 9 Symptoms-contrast
Morris-Stiff et al12 1997 10 0.5–15 Contrast manometry HM 8 Symptoms-weight
Patti et al.19 2001 13 6–17 Contrast-manometry-scope HM 9.5 Symptoms
Hussain et al20 2002 33 0.4–16 Contrast manometry scope HM-EOD- 4.7 Symptoms-weight
Botox
Upadhyaya et al21 2002 12 0.5–14 Contrast EOD 3.5 Symptoms
Garzi et al22 2007 14 3.5–16 Contrast-Manometry-Scope HM NS Symptoms-manometry-weight
23
Paidas et al 2007 14 12–18 Contrast-manometry-scope HM 3 Symptoms
Pastor et al24 2009 40 4–8 Contrast-manometry-scope HM-EOD 6.2 Symptoms
Vaos et al25 2008 15 6–13 Contrast-manometry-scope HM 15 Symptoms-manometry-
contrast
Askegard- 2009 26 4–18 Contrast-manometry HM 1.7 Symptoms
Giesmann et al26
Corda et al27 2010 20 5–15 Contrast-manometry-scope HM 5 Symptoms
Tannuri et al28 2010 15 9–17 Contrast-manometry-scope HM 2.69 Symptoms
29
Zhang et al 2009 13 3–14.5 Contrast manometry scope HM-EOD- 1.26 Symptoms
Nifedipine
Jung et al30 2010 22 0.3–17 Contrast-manometry HM-EOD 2 Symptoms-weight
Di Nardo et al31 2012 24 5–17 Contrast-manometry EOD 6 Contrast-manometry- Eckardt
Esposito et al32 2013 31 5–15 Contrast-manometry-scope HM NS Symptoms
Caldaro et al33 2015 18 2–17 Contrast-manometry-scope HM-POEM 2.58 Manometry- Eckardt
Chen et al34 2015 27 6–17 Contrast-manometry-scope POEM 2 Manometry-Eckardt
Meyer et al35 2017 42 7–13 Contrast Manometry HM-EOD- 4.4 Symptoms
Botox-
Nifedipine
Nabi et al36 2016 15 9–18 Contrast-manometry-scope POEM 1.16 Symptoms-manometry-
contrast-weight-Eckardt-scope
Smits et al37 2016 86 9–14 Contrast manometry scope HM-EOD 3.9 Symptoms
38
Tan et al 2016 21 6–17 Contrast-manometry-scope POEM-EOD NS Eckardt
Zagory et al39 2016 26 0.5–17 N/S HM-EOD 2.5 Symptoms
40
Grabowski et al 2017 11 6–17 Contrast Manometry scope HM 2.5 Symptoms-weight
Stavropoulos 2017 10 10–17 N/S POEM 1.25 Manometry-contrast-
et al41 Eckardt- PH-metry
Saliakellis et al42 2017 48 3–17 Contrast-manometry-scope-biopsy HM-EOD 3 Eckardt
Duggan et al43 2019 31 0.25–16 Contrast-manometry-scope HM 1.5 Symptoms
44
Miao et al 2018 21 0.9–18 Contrast-manometry-scope POEM 1.1 Symptoms-contrast-Eckardt
Abbreviations: Eckardt, Eckardt score; EOD, endoscopic esophageal dilatation; HM, Heller myotomy; NS, not specified; POEM, peroral esoph-
agomyotomy; Scope, esophagoscopy.
nifedipine (3%). Mean follow-up was 43.7 months (12–180). jective measures were only used in 13 studies (39.4%), which
Outcome measures were heterogeneous across studies and lead to variable definitions of operative success. However,
included symptomatic improvement, Eckardt score, weight analysis of reported success showed an average success of
gain, postoperative manometry, and contrast swallow. Ob- 77.9% for HM (p ¼ 1.79 107), 44.9% for EOD (p ¼ 0.24), and

99.3% for POEM (p ¼ 0.001). Complications were reported in was found to have lower complication rates (5%), it had less
12.8% of patients who had HM, 5% for EOD, and 24.4% for success (44.9%) in the short and medium term and most cases
POEM. Further surgical interventions were required for required either repeated EOD or combination with other
10.9% of HM, 62.3% of EOD, and 0.01% of POEM patient treatment modalities. One paper described a favorable out-
groups. come after the first EOD in older children (> 12 years) with
higher baseline LES that had a significant decrease in LES
pressure after the dilatation.12 This can make a good selec-
Discussion
tion group for which EOD can be considered as a primary
When children present with symptoms suggestive of OA, a treatment.
set of diagnostic tools is used to help reach the diagnosis. In the POEM group (►Table 4), there were six studies
Most studies (84.8%) depended on contrast esophagogram included in the analysis. Two were comparative studies, one
and manometric studies to confirm the diagnosis. Manome- comparing it to HM and another to EOD. Ninety-four patients
try is a valuable tool in diagnosis of OA, since it also provides underwent a POEM procedure, of which 20 patients (21.3%)
information about the LES pressure, which was seen to underwent prior treatments. Patients’ average age at diag-
influence the treatment choice and outcome in one study.12 nosis was 13.75 (11 months to 18 years). Only one patient
Published studies for treatment of OA in children had was diagnosed with astrocytoma as a comorbidity. This
multiple limitations. Numbers of patients in each study were group showed the best average objective success rate
relatively small due to rarity of the disease. Outcome meas- (99.3%) in comparison to the previous two groups. However,
ures described were not unified with most studies describing it is a newly introduced technique which makes widely
subjective measures. Some patients presented with surgical adopting it a challenge. It requires specific equipment and
or medical treatments prior to the primary treatment de- advanced endoscopic skills. Although success rates were
scribed which inevitably affected the overall outcome. There relatively higher for the POEM group, only short-term fol-
is no clear definition of postoperative success, which made low-up data was available and medium- and long-term
the comparison between treatment modalities difficult. follow-up data are still not available to judge how it compares
Many patients had symptomatic relief after a combination to other techniques. The most encountered complication was
of treatment methods, these combinations have not been mucosal tears (9.7%) and it was either managed endoscopi-
assessed separately for their effect on outcomes. Follow-up cally if discovered during the operation or otherwise treated
was acceptable for the medium term, but data are still conservatively if discovered postoperatively.
lacking for long-term outcomes for children. Patients’ characteristics were heterogeneous across all
In the HM group (►Table 2), there were 25 studies three treatment groups when it came to age at diagnosis,
analyzed, of which 11 compared it to other treatment comorbidities, preceding treatments, and presenting symp-
modalities, and only 1 study was a prospective study. Note toms. No attribution bias could be identified since there were
that 430 patients underwent HM across studies, of which no overriding characteristics in individual treatment groups
130 patients underwent treatments prior to surgery (30%). (►Table 5).
Patients’ average age at diagnosis was 10.8 (15 days to 17.6 The American Collage of Gastroenterology guidelines for
years). Eighty-one patients (18.8%) presented with other diagnosis and management of OA in adults5 included inter-
comorbidities (e.g., AAA syndrome, trisomy 21, etc.). Vari- esting points that can be useful in pediatric practice. They
able surgical techniques were described, including minimal- highlighted the role of manometry as a diagnostic tool with
ly invasive surgeries (laparoscopic or thoracoscopic), limited superior sensitivity over contrast esophagogram and endos-
or generous, and modified HM with or without various copy. They advocate a tailored approach to treatment of OA
antireflux procedures. The added benefit of antireflux pro- focused on symptomatic relief, improvement of gastric emp-
cedures has shown better control of gastroesophageal reflux tying, and prevention of further interventions. Laparoscopic
after a generous myotomy, and a decrease in medium term HM with antireflux procedures came superior to single EOD
morbidity. There did not appear to be one surgical technique in short- and medium-term follow-up (89.3% vs. 56.3%).
superior to others. However, described success rates were However, results were comparable when compared with
seen to be superior (77.9%) to patients managed with medi- serial repeated EOD.
cal treatment, botulinum injections, or repeated EOD. There
was a higher tendency for mucosal perforations for patients
Conclusion
who previously had EOD, and the overall reported compli-
cations were seen to be higher in this group (12.8%), making In conclusion and in light of the evidence scrutinized in the
it a more risky but effective procedure on the medium term. context of this systematic review, we cannot confidently
In the EOD group (►Table 3), 11 out of a total of 14 studies declare one treatment modality as the standard of treatment
compared EOD to other treatments. Two studies were pro- of OA in children. HM appears to be a more successful
spective. Note that 218 patients underwent single or repeat- intervention than EOD. POEM is a promising intervention,
ed EOD as their main treatment, of which 20 patients (9.1%) but its sustainability is yet to be researched. A multicentric
received previous treatments. Patients’ average age at diag- prospective randomized trial with long-term follow-up is
nosis was 11.8 (2 months to 17.4 years). Fifty patients (22.9%) needed with clearer diagnostic criteria, uniform outcome
presented with comorbidities. Even though this technique measures, and a clearer definition of success.

Table 2 Results of Heller myotomy as treatment for esophageal achalasia
Author Year N Study type Prior treatment Outcome measure Success Subsequent treatment Complications
Azizkhan et al13 1980 11 Retrospective 100% EOD Symptoms-contrast-weight 81.8% 9% EOD þ Repeat HM 9% Hemorrhage – 9% mild GOR
Vane et al15 1988 21 Retrospective 19% EOD Symptoms 85.7% 4.7% EOD – 19% HM 9.5% Perforation
Nihoul-Fékété et al16 1989 35 Retrospective None Symptoms-contrast-weight 97.1% 3% Repeat HM 8.5% Perforation
Emblem et al17 1993 12 Retrospective 58% EOD – 33% Symptoms-contrast- 75% 42% EOD 16.6% Food bolus – 8% lung
Nifedipine – 8% HM weight-scope collapse – 8% gas bloat
Lelli et al18 1997 19 Retrospective N/S Symptoms 89.5% 100% EOD None
12
Morris-Stiff et al 1997 10 Retrospective N/S Symptoms-weight 80% N/S 30% minor complications
Patti et al.19 2001 13 Retrospective 15% EOD – 8% Botox Symptoms 100% N/S None
Hussain et al20 2002 17 Retrospective 65% EOD Symptoms-weight 88.2% 7% Botox 17% GOR
22
Garzi et al 2007 14 Retrospective None Symptoms-manometry- 71.1% 7% Repeat HM 7% Perforation
weight
Paidas et al23 2007 14 Retrospective 7% Botox – 14% EOD – Symptoms 78.6% 14% Repeat HM 7% Perforation
28% EOD þ Botox
Pastor et al24 2009 10 Retrospective 20% HM Symptoms 60% 20% EOD – 10% esophagectomy – 10% GOR – 20% perforation
5% antireflux surgery
Vaos et al25 2008 15 Retrospective N/S Symptoms-manometry- 93.3% 6.7% Repeat HM 13% Minor complications
contrast
Askegard-Giesmann et al26 2009 9 Retrospective N/S Symptoms 88.9% 11% Repeat HM N/S
Corda et al27 2010 26 Retrospective 53.8% EOD – Symptoms 69.2% 11.5% Repeat HM – 7.7% Perforation –
23% Botox 11.5% EOD Botox 3.8% aspiration
Tannuri et al28 2010 20 Retrospective 30% EOD Symptoms 75% 10% EOD – 10% Repeat HM 15% Perforation
Zhang et al29 2009 15 Retrospective 33% EOD Symptoms 86.7% 6.6% Botox None
30
Jung et al 2010 15 Prospective 13% EOD Symptoms-weight 66.7% 13% EOD – 20% HM 6.6% Aspiration
Esposito et al32 2013 31 Retrospective 6.5% Nifedipine Symptoms 96.8% 3.2% Repeat HM – 6.4% EOD 9.6% Perforation
Caldaro et al33 2015 9 Retrospective 33% EOD – 11% CCB Manometry-Eckardt 100% 22% EOD 11% Esophageal perforation
Meyer et al35 2017 17 Retrospective N/S Symptoms 35% 35% EOD – 29% Botox – 18% Gastric perforation
29% Repeat HM
Smits et al37 2016 18 Retrospective N/S Symptoms 77.8% N/S 16.7% Perforation – total 55.6%
Zagory et al39 2016 9 Retrospective N/S Symptoms 66.7% 22% Repeat HM – 11% EOD 11% Perforation
Grabowski et al40 2017 11 Retrospective 36% EOD Symptoms-weight 54.5% 27% EOD 16% Perforation
Surgical Management of Esophageal Achalasia in Pediatrics
Saliakellis et al42 2017 28 Retrospective CCB/Nifedipine n ¼ ? Eckardt 60.7% 35.7% EOD – 3.6% Repeat HM 5.5% Esophageal perforation
Duggan et al43 2019 31 Retrospective 42% EOD Symptoms 71% 6% EOD – 3% Repeat HM 7% GOR – 6.5% perforation
Total ¼ 25 430 1 prospective 72% objective 77.9% 12.8%
(35.2–100%)
Abbreviations: CCB, calcium channel blocker; Eckardt, Eckardt score; EOD, endoscopic esophageal dilatation; GOR, gastroesophageal reflux; HM, Heller myotomy; N/S, not specified; Scope, esophagoscopy.

Goneidy et al.
Table 3 Results of endoscopic esophageal dilatation as treatment for esophageal achalasia
Author Year N Study type Prior treatment Outcome measure Success Subsequent treatment Complications
Azizkhan et al13 1980 8 Retrospective None Symptoms-contrast-weight 25 50% EOD 12.5% Aspiration – 25% GOR
Boyle et al 14 1981 10 Retrospective N/S Symptoms-weight 40 20% EOD – 20% HM 10% Sever pain – no perforation
17
Emblem et al 1993 2 Retrospective N/S Symptoms-contrast-weight-scope 50 N/S N/S
Upadhyaya et al 21 2002 12 Retrospective N/S Symptoms 83.3 17% EOD None
Hussain et al20 2002 9 Retrospective N/S Symptoms-weight 0 100% HM N/S
24
Pastor et al 2009 30 Retrospective None Symptoms 33.3 53% EOD – 30% HM 3.3% GOR – 7% aspiration

Zhang et al29 2009 2 Retrospective N/S Symptoms 100 50% Nifedipine N/S
Jung et al30 2010 7 Prospective 28.5% HM Symptoms-weight 57.1 22% Rpt EOD None
31
Di Nardo et al 2012 24 Prospective N/S Contrast-manometry-Eckardt 66.7 12.5% HM – 20.8% Repeat EOD 11% Pain
Meyer et al35 2017 3 Retrospective N/S Symptoms 33.3 66% Botox N/S
Smits et al37 2016 68 Retrospective N/S Symptoms 10.3 22% EOD 1.5% Perforation
38
Tan et al 2016 9 Retrospective N/S Eckardt 100 ND None
Zagory et al39 2016 14 Retrospective N/S Symptoms 0 71% HM – 29% ED – 21% Botox N/S
Saliakellis et al42 2017 20 Retrospective CCB/Nifedipine n ¼ ? Eckardt 30 25% EOD – 60% HM 5% Esophageal perforation
Total ¼ 14 218 2 prospective 36% objective 45% (0–100%) 5%
Abbreviations: CCB, calcium channel blocker; Eckardt, Eckardt score; EOD, endoscopic esophageal dilatation; GOR, gastroesophageal reflux; HM, Heller myotomy; N/S, not specified; Scope, esophagoscopy.
Surgical Management of Esophageal Achalasia in Pediatrics
Table 4 Results of POEM as treatment for esophageal achalasia
Author Year N Study type Prior treatment Outcome measure Success Subsequent Complications
treatment
Caldaro et al33 2015 9 Retrospective 1 EOD Manometry-Eckardt 100 None 11% Mucosal tear – 11% pneumoperitoneum
Chen et al34 2015 27 Prospective 19% EOD – 4% Botox – 4% Stent Manometry-Eckardt 96.2 N/S 18.5% Mucosal tear – 3.7% pneumothorax
Goneidy et al.
Tan et al38 2016 12 Retrospective N/S Eckardt 100 N/S 16.7% GOR – 8.3% subcutaneous emphysema
36
Nabi et al 2016 15 Retrospective 40% EOD – 6.7% HM Symptoms-manometry- 100 None 20% GOR – 6.7% perforation
contrast-weight-Eckardt-scope
Stavropoulos et al41 2017 10 Retrospective 20% HM – 20% EOD Manometry-contrast-Eckardt-PH-metry 100 None None
Miao et al44 2018 21 Retrospective 4.7% EOD Symptoms-contrast-Eckardt 100 None 29% GOR – 9.5% perforation
Total ¼ 6 93 1 Prospective 100% Objective 99.3% (96.2–100%) 24.4%
Abbreviations: Eckardt, Eckardt score; EOD, endoscopic esophageal dilatation; GOR, gastroesophageal reflux; HM, Heller myotomy; N/S, not specified; POEM, peroral esophageal myotomy; Scope,
esophagoscopy.
Table 5 Comparison between three patient groups’ treatment 16 Nihoul-Fékété C, Bawab F, Lortat-Jacob S, Arhan P, Pellerin D.
modalities Achalasia of the esophagus in childhood: surgical treatment in 35
cases with special reference to familial cases and glucocorticoid
Treatment Heller Endoscopic POEM deficiency association. J Pediatr Surg 1989;24(10):1060–1063
myotomy esophageal 17 Emblem R, Stringer MD, Hall CM, Spitz L. Current results of surgery
dilatation for achalasia of the cardia. Arch Dis Child 1993;68(06):749–751
18 Lelli JL Jr, Drongowski RA, Coran AG. Efficacy of the transthoracic
Total number 430 218 94
modified Heller myotomy in children with achalasia–a 21-year
of patients
experience. J Pediatr Surg 1997;32(02):338–341
Average age 10.8 years 11.8 years 13.75 years 19 Patti MG, Albanese CT, Holcomb GW III, et al. Laparoscopic Heller
at treatment myotomy and Dor fundoplication for esophageal achalasia in
Average 3 years 3.5 years 1.3 years children. J Pediatr Surg 2001;36(08):1248–1251
follow-up 20 Hussain SZ, Thomas R, Tolia V. A review of achalasia in 33 children.
Dig Dis Sci 2002;47(11):2538–2543
Prior treatments 30% 9.1% 21.3%
21 Upadhyaya M, Fataar S, Sajwany MJ. Achalasia of the cardia:
Comorbidities 18.8% 22.9% 0.01% experience with hydrostatic balloon dilatation in children.
Pediatr Radiol 2002;32(06):409–412
Complications 12.8% 5% 24.4%
22 Garzi A, Valla JS, Molinaro F, Amato G, Messina M. Minimally
Success rate 77.9% 44.9% 99.3% invasive surgery for achalasia: combined experience of two Euro-
Further surgical 10.9% 62.3% 0.01% pean centers. J Pediatr Gastroenterol Nutr 2007;44(05):587–591
intervention 23 Paidas C, Cowgill SM, Boyle R, Al-Saadi S, Villadolid D, Rosemurgy
AS. Laparoscopic Heller myotomy with anterior fundoplication
Abbreviation: POEM, peroral esophageal myotomy. ameliorates symptoms of achalasia in pediatric patients. J Am Coll
Surg 2007;204(05):977–983
24 Pastor AC, Mills J, Marcon MA, Himidan S, Kim PC. A single center
Conflict of Interest 26-year experience with treatment of esophageal achalasia: is
None declared. there an optimal method? J Pediatr Surg 2009;44(07):1349–1354
25 Vaos G, Demetriou L, Velaoras C, Skondras C. Evaluating long-term
results of modified Heller limited esophagomyotomy in children
References with esophageal achalasia. J Pediatr Surg 2008;43(07):1262–1269
1 Boeckxstaens GE, Zaninotto G, Richter JE. Achalasia. Lancet 2014; 26 Askegard-Giesmann JR, Grams JM, Hanna AM, Iqbal CW, Teh S,
383(9911):83–93 Moir CR. Minimally invasive Heller’s myotomy in children: safe
2 Richter JE. Oesophageal motility disorders. Lancet 2001;358 and effective. J Pediatr Surg 2009;44(05):909–911
(9284):823–828 27 Corda L, Pacilli M, Clarke S, Fell JM, Rawat D, Haddad M. Laparo-
3 Marlais M, Fishman JR, Fell JM, Haddad MJ, Rawat DJ. UK incidence scopic oesophageal cardiomyotomy without fundoplication in
of achalasia: an 11-year national epidemiological study. Arch Dis children with achalasia: a 10-year experience: a retrospective
Child 2011;96(02):192–194 review of the results of laparoscopic oesophageal cardiomyotomy
4 Mayberry JF, Mayell MJ. Epidemiological study of achalasia in without an anti-reflux procedure in children with achalasia. Surg
children. Gut 1988;29(01):90–93 Endosc 2010;24(01):40–44
5 Vaezi MF, Pandolfino JE, Vela MF. ACG clinical guideline: diagnosis 28 Tannuri AC, Tannuri U, Velhote MC, Romão RL. Laparoscopic
and management of achalasia. Am J Gastroenterol 2013;108(08): extended cardiomyotomy in children: an effective procedure
1238–1249 for the treatment of esophageal achalasia. J Pediatr Surg 2010;
6 Maksimak M, Perlmutter DH, Winter HS. The use of nifedipine for 45(07):1463–1466
the treatment of achalasia in children. J Pediatr Gastroenterol 29 Zhang Y, Xu CD, Zaouche A, Cai W. Diagnosis and management of
Nutr 1986;5(06):883–886 esophageal achalasia in children: analysis of 13 cases. World J
7 Ip KS, Cameron DJ, Catto-Smith AG, Hardikar W. Botulinum toxin for Pediatr 2009;5(01):56–59
achalasia in children. J Gastroenterol Hepatol 2000;15(10):1100–1104 30 Jung C, Michaud L, Mougenot JF, et al. Treatments for pediatric
8 Gariepy CE, Mousa H. Clinical management of motility disorders achalasia: Heller myotomy or pneumatic dilatation? Gastroen-
in children. Semin Pediatr Surg 2009;18(04):224–238 terol Clin Biol 2010;34(03):202–208
9 Familiari P, Marchese M, Gigante G, et al. Peroral endoscopic 31 Di Nardo G, Rossi P, Oliva S, et al. Pneumatic balloon dilation in
myotomy for the treatment of achalasia in children. J Pediatr pediatric achalasia: efficacy and factors predicting outcome at a
Gastroenterol Nutr 2013;57(06):794–797 single tertiary pediatric gastroenterology center. Gastrointest
10 Greenhalgh T. Assessing methodological quality. In: Greenhalgh T, ed. Endosc 2012;76(05):927–932
How to Read a Paper: The Basics of Evidence-Based Medicine. 5th ed. 32 Esposito C, Riccipetitoni G, Chiarenza SF, et al. Long-term results
Hoboken, NJ: Wiley Blackwell; 1997:45–59 of laparoscopic treatment of esophageal achalasia in children: a
11 Downs SH, Black N. The feasibility of creating a checklist for the multicentric survey. J Laparoendosc Adv Surg Tech A 2013;23
assessment of the methodological quality both of randomised and (11):955–959
non-randomised studies of health care interventions. J Epidemiol 33 Caldaro T, Familiari P, Romeo EF, et al. Treatment of esophageal
Community Health 1998;52(06):377–384 achalasia in children: Today and tomorrow. J Pediatr Surg 2015;50
12 Morris-Stiff G, Khan R, Foster ME, Lari J. Long-term results of surgery (05):726–730
for childhood achalasia. Ann R Coll Surg Engl 1997;79(06):432–434 34 Chen WF, Li QL, Zhou PH, et al. Long-term outcomes of peroral
13 Azizkhan RG, Tapper D, Eraklis A. Achalasia in childhood: a 20-year endoscopic myotomy for achalasia in pediatric patients: a prospec-
experience. J Pediatr Surg 1980;15(04):452–456 tive, single-center study. Gastrointest Endosc 2015;81(01):91–100
14 Boyle JT, Cohen S, Watkins JB. Successful treatment of achalasia in 35 Meyer A, Catto-Smith A, Crameri J, et al. Achalasia: outcome in
childhood by pneumatic dilatation. J Pediatr 1981;99(01):35–40 children. J Gastroenterol Hepatol 2017;32(02):395–400
15 Vane DW, Cosby K, West K, Grosfeld JL. Late results following 36 Nabi Z, Ramchandani M, Reddy DN, et al. Per oral endoscopic
esophagomyotomy in children with achalasia. J Pediatr Surg myotomy in children with achalasia cardia. J Neurogastroenterol
1988;23(06):515–519 Motil 2016;22(04):613–619

37 Smits M, van Lennep M, Vrijlandt R, et al. Pediatric achalasia in the 41 Stavropoulos SN, Sosulski AB, Modayil RJ, et al. Sa2075 use of
Netherlands: incidence, clinical course, and quality of life. J Pediatr peroral endoscopic myotomy (POEM) in pediatric patients as a
2016;169:110–115 primary or rescue therapy for achalasia. Gastrointest Endosc
38 Tan Y, Zhu H, Li C, Chu Y, Huo J, Liu D. Comparison of peroral 2017;85:AB285–AB286
endoscopic myotomy and endoscopic balloon dilation for primary 42 Saliakellis E, Thapar N, Roebuck D, et al. Long-term outcomes of
treatment of pediatric achalasia. J Pediatr Surg 2016;51(10): Heller’s myotomy and balloon dilatation in childhood achalasia.
1613–1618 Eur J Pediatr 2017;176(07):899–907
39 Zagory JA, Golden JM, Demeter NE, Nguyen Y, Ford HR, Nguyen 43 Duggan EM, Nurko S, Smithers CJ, Rodriguez L, Fox VL, Fishman SJ.
NX. Heller myotomy is superior to balloon dilatation or botulinum Thoracoscopic esophagomyotomy for achalasia in the pediatric
injection in children with achalasia: a two-center review. population: a retrospective cohort study. J Pediatr Surg 2019;54
J Laparoendosc Adv Surg Tech A 2016;26(06):483–487 (03):572–576
40 Grabowski A, Korlacki W, Pasierbek M, Pułtorak R, Achtelik F, 44 Miao S, Wu J, Lu J, et al. Peroral endoscopic myotomy in children
Ilewicz M. Pediatric achalasia. Single-center study of interven- with achalasia: a relatively long-term single-center study.
tional treatment. Prz Gastroenterol 2017;12(02):98–104 J Pediatr Gastroenterol Nutr 2018;66(02):257–262

Original Article
Asthma Medication Use in Congenital Diaphragmatic

Hernia Survivors: A Retrospective Population Level
Data Analysis
Matthew Levesque1 Suyin A. Lum Min1 Melanie I. Morris1 Anna C. Shawyer1 Richard Keijzer1
1 Department of Surgery, Division of Pediatric Surgery, University of Address for correspondence Richard Keijzer, MD, MSc, PhD,
Manitoba and Children’s Hospital Research Institute of Manitoba, Department of Surgery, Division of Pediatric Surgery, University of
Winnipeg, Manitoba, Canada Manitoba and Children’s Hospital Research Institute of Manitoba,
Winnipeg, Ae402-820 Sherbrook Street, Winnipeg, Manitoba R3A1S1,
Eur J Pediatr Surg Canada (e-mail: [email protected]).
Abstract Introduction The purpose of this study was to determine if congenital diaphragmatic
hernia (CDH) survivors had worse long-term respiratory outcomes compared with age-
matched controls, as measured by inhaled bronchodilator use, inhaled steroid use, and
asthma-related physician visits.
Materials and Methods We performed a retrospective case-control study of infants with
isolated CDH from 1991 to 2013. The primary outcome measures were inhaled broncho-
dilator prescriptions, inhaled steroid prescriptions, and asthma-related physician visits
between 0 and 5 years of age and between 5 and 10 years of age. Subgroup analysis
compared the same outcomes for CDH patients grouped by: birth weight, gestational age,
side of defect, defect size, liver herniation, hernia sac, and pulmonary hypertension.
Results Fifty-six patients with CDH and 753 age-matched controls met the inclusion
criteria for the 0 to 5 years of age analysis. Between 0 and 5 years of age, more CDH
survivors were prescribed an inhaled bronchodilator (odds ratio [OR] ¼ 2.47[1.38–
4.48], p ¼ 0.001) and inhaled steroid (OR ¼ 2.03[1.07–3.74], p ¼ 0.03), and had an
asthma-related physician visit (OR ¼ 1.92[1.00–3.56], p ¼ 0.04). Thirty-eight cases and
491 controls met the inclusion criteria for the 5 to 10 years of age analysis. Between 5
and 10 years of age, CDH survivors were not more likely to be prescribed inhaled
bronchodilators, inhaled steroids, or have an asthma-related physician visit. Among the
CDH patients, we did not find a clinical characteristic associated with increased inhaled
Keywords bronchodilator or steroid prescriptions at any age.
► congenital Conclusion A history of CDH is associated with higher rates of inhaled bronchodilator
diaphragmatic hernia prescriptions, inhaled steroid prescriptions, and asthma-related physician visits from 0
► long-term outcomes to 5 years of age compared with age-matched controls. However, this difference
► respiratory outcomes resolves by 5 to 10 years of age.
Introduction in Canada is approximately 3.38/10,000 births1 and the mor-

tality rate of CDH is approximately 20%.2,3 Numerous studies
Congenital diaphragmatic hernia (CDH) is a birth defect char- have found that children with CDH have pulmonary, cardiovas-
acterized by a hole in the diaphragm. The incomplete dia- cular, gastrointestinal, neurocognitive, and musculoskeletal
phragm allows viscera from the abdomen to herniate into the complications.4–10 Most of these studies focus on early out-
thoracic cavity while the fetus is in utero. The incidence of CDH comes occurring within the first 1 to 2 years of life. There is less

September 2, 2019
Congenital Diaphragmatic Hernia Survivors Levesque et al.
research on outcomes of CDH survivors beyond 2 years of age. inhaled steroids based on the Anatomical Therapeutic Chemi-
As neonatal management techniques improve and more chil- cal classification code. An Anatomical Therapeutic Chemical
dren with CDH survive through childhood, understanding long- classification code exists for every medication and the code is
term morbidity is important. organized based on the organ system and mechanism of
The purpose of this study was to determine if CDH survivors action. We defined bronchodilators as any medication that
had worse long-term respiratory outcomes compared with age- fell under the “inhaled adrenergic” (R03A) or “inhaled anti-
matched controls as measured by inhaled bronchodilator cholinergic” (R03BB) categories, and steroids as any medica-
prescriptions, inhaled steroid prescriptions, and asthma-related tion that fell under the “inhaled glucocorticoids” (R03BA)
physician visits. Additionally, we aimed to determine if previ- category. The Medical Claims/Medical Services is an adminis-
ously established unfavorable CDH clinical characteristics were trative dataset of billings through which physicians receive
associated with worse long-term respiratory outcomes. The compensation for services rendered. This dataset includes
clinical characteristics analyzed were birthweight, gestational information on every primary care physician visit in Manitoba,
age, right-sided defect, large defect size, liver herniation, including the diagnosis made at the visit. We used this dataset
absence of a hernia sac, and pulmonary hypertension. to determine which individuals in the cohort had an asthma-
related physician visit. The diagnosis at each visit was coded
using the International Classification of Disease, ninth revi-
sion. We defined an asthma-related physician visit as any visit
Study Population with a diagnosis code of “asthma” (493).
The CDH case cohort was derived from the Winnipeg Surgical
Database of Outcomes and Management (WiSDOM). The Statistical Analysis
WiSDOM database contains demographic information and The primary aim was to determine if proportionately more
clinical details of all the children with CDH treated at Health CDH cases than controls were prescribed inhaled broncho-
Sciences Centre and/or St. Boniface Hospital since 1991. Heath dilators, inhaled steroids, or had an asthma-related physician
Sciences Centre and St. Boniface Hospital are the only sites to visit between 0 and 5 years of age and between 5 and 10 years
provide pediatric surgical treatment to the children born with of age. For the 0 to 5 years of age analysis, we included all the
CDH in Manitoba. This allows for the continuity of patient care children in the cohort born between 1991 and 2013 and for
and the collection of longitudinal data. Ethical approval was the 5 to 10 years of age analysis we included all the children
obtained to record and analyze the data from this database in the cohort born between 1991 and 2008. These outcomes
(HS19268 [H2016:014]). The inclusion criteria consisted of all were analyzed using Fisher’s exact tests.
the children born in Manitoba with a Bochdalek-type CDH The secondary aim was to determine how clinical char-
defect between 1991 and 2013. The control cohort was acteristics influence the number of CDH cases who were
generated using the Manitoba Health Insurance Registry; for prescribed inhaled bronchodilators, inhaled steroids, or had
each CDH case, we identified all the children born in Manitoba an asthma-related physician. Similar to the primary aim, we
on the same date. Of the children born on the same date, 10 used the 0 to 5 years of age time frame and 5 to 10 years of age
were randomly selected to be included in the control group. time frame. Clinical characteristics considered were birth-
Once the cohort was established, cases and controls were weight, gestational age, right-sided defect, large defect size,
excluded if their provincial health care coverage was incom- liver herniation, absence of a hernia sac, and pulmonary
plete for either time frame as indicated in the Manitoba Health hypertension. Pulmonary hypertension severity was deter-
Insurance Registry. Health care in Manitoba is universal and mined as mild or severe by echocardiogram. The criteria for
publicly funded; it is relinquished if a person moves out of the severe pulmonary hypertension included the presence of
province or dies. either: bidirectional or right-to-left shunt through a patent
ductus arteriosus, or septal flattening. Defect size was cate-
Data Reviewed gorized into the A-D scale defined by the Congenital Dia-
There were three outcome measures for this study; inhaled phragmatic Hernia Study Group13; categories A and B were
bronchodilator use, inhaled steroid use, and asthma-related classified as small, and categories C and D were classified as
physician visits. The outcomes were assessed using data from large. Logistic regression was used for these analyses.
the Manitoba Centre for Health Policy, a repository of datasets Data analysis was performed using R statistical software.
that facilitates population-based research into factors that For all analyses, we used two-tailed tests and p-values <0.05
affect health.11,12 The datasets within the Manitoba Centre were considered statistically significant.
for Health Policy that we used were the Drug Program Infor-
mation Network and Medical Claims/Medical Services. The Drug
Program Information Network is an administrative dataset of Results
prescription drug dispensation in Manitoba; all prescription
Primary Aim—Medication Use and Physician Visits
drugs dispensed within Manitoba are recorded in this dataset
for Cases versus Controls
and linked to the individual for whom the prescription was
written. We used this dataset to determine which individuals 0 to 5 Years of Age
in the cohort were prescribed inhaled bronchodilators and There were 82 cases and 818 date of birth-matched controls
inhaled steroids. We identified inhaled bronchodilators and born between 1991 and 2013; for two cases only nine

controls were matched. Of the 900 children, 91 were exclud- asthma-related physician visits between 0 and 5 years of age.
ed because they did not have health coverage for 0 to 5 years We found that a lower birthweight and lower gestational age
of age time frame; 26 cases (19 died; 7 emigrated) and 65 were associated with increased rates of both inhaled steroid
controls (9 died; 56 emigrated). The remaining 809 individ- prescriptions (p ¼ 0.024 and 0.013, respectively) and a lower
uals, 56 cases and 753 controls, constituted the study popu- birthweight was associated with more asthma-related physi-
lation for assessment of the 0 to 5 years of age outcomes. cian visits (p ¼ 0.032). Right-sided defect, large defect size, liver
We found that more CDH survivors were prescribed an herniation, absence of a hernia sac, and severe pulmonary
inhaled bronchodilator (OR ¼ 2.47, 95% CI ¼ [1.38–4.48], hypertension each were not found to be associated with
p ¼ 0.001), inhaled steroid (OR ¼ 2.03[1.07–3.74], p ¼ 0.026), increased inhaled bronchodilator prescriptions, inhaled ste-
and had an asthma-related physician visit (OR ¼ 1.92[1.00– roid prescriptions, or asthma-related physician visits.
3.56], p ¼ 0.038) than controls. Additionally, CDH survivors
had a greater mean length of inhaled bronchodilator prescrip- 5 to 10 Years of Age
tions (108 vs. 50.3 days, d ¼ 57.8[8.9–106] days) and a greater For the 38 cases from the primary 5 to 10 years of age
mean length of inhaled steroid prescriptions (114 vs. 48.3 days, analysis, we did not find any of the clinical characteristics
d ¼ 66.0[3.8–128] days). (lower birthweight, lower gestational age, right-sided defect,
large defect size, liver herniation, absence of a hernia sac, or
5 to 10 Years of Age severe pulmonary hypertension) to be associated with
A total of 627 children were born between 1991 and 2008; 57 higher rates of inhaled bronchodilator prescriptions, inhaled
cases with CDH and 570 date of birth-matched controls. Of the steroid prescriptions, or asthma-related physician visits. The
627 children, 98 were excluded because they did not have results of the secondary analysis between 0 and 5 years of age
complete health coverage from 5 to 10 years of age; 19 cases and 5 and 10 years of age are summarized in ►Table 2.
(13 died; 6 emigrated) and 79 controls (8 died; 71 emigrated).
The remaining 529 individuals, 38 cases and 491 controls,
Discussion
constituted the study population for assessment of the 5 to
10 years of age outcomes. This study used population datasets to perform a virtual
Comparing CDH survivors to controls, we found no differ- follow-up of CDH patients to assess long-term respiratory
ences in the number of children prescribed an inhaled bron- function as suggested by two surrogate markers: prescrip-
chodilator (OR ¼ 1.49[0.622–3.29], p ¼ 0.295), inhaled steroid tion drug use and asthma-related physician visits. We found
(OR ¼ 1.56[0.56–3.82], p ¼ 0.316), or had an asthma-related that CDH survivors were prescribed inhaled bronchodilators,
physician visit (OR ¼ 1.78[0.74–3.96], p ¼ 0.179). Additionally, inhaled steroids, and had asthma-related physician visits
no differences were found in the mean length of inhaled more often than children without a history of CDH, up to
bronchodilator prescriptions (42.3 vs. 21.3 days, d ¼ 21.1[(-) 5 years of age. We did not find any differences in the rates of
19.0–61.1] days) or inhaled steroid prescriptions (60.5 vs. 24.3 inhaled bronchodilator prescriptions, inhaled steroid pre-
days, d ¼ 36.2[(-)19.5–91.9] days). The results of the primary scriptions, or asthma-related physician visits between 5 and
analysis between 0 and 5 years of age and 5 and 10 years of age 10 years of age.
are summarized in ►Table 1. To our knowledge, long-term respiratory outcomes have
not been extensively reported for CDH survivors. El Cheha-
deh et al reported that at a mean of 6.6 years of age,
Secondary Aim—Medication Use and Physician Visits
respiratory symptoms occurred in 30% of 56 patients, and
in Cases Grouped by Clinical Characteristics
abnormal pulmonary function tests were common among
0 to 5 Years of Age those tested.14 Koziarkiewicz et al reported recurrent respi-
Using the 56 cases from the primary 0 to 5 years of age analysis, ratory system infections in 34% of 50 patients followed long
we performed a logistic regression to determine if any clinical term.15 Kamata et al reported on 33 children assessed at
characteristics were associated with increased rates of inhaled 4.1 2.5 years of age noting that frequent respiratory tract
bronchodilator prescriptions, inhaled steroid prescriptions, or infections were common; they also noted that the frequency
Table 1 Results of the primary analysis between 0 and 5 years of age and 5 and 10 years of age
Year Outcome CDH Control OR 95% CI p-Value

0–5 Inhaled bronchodilator 32/56 (57.1%) 264/753 (35.1%) 2.467 1.376–4.476 0.001a
Inhaled steroid 19/56 (33.9%) 152/753 (20.2%) 2.028 1.070–3.738 0.026a
Asthma-related physician visit 18/56 (32.1%) 149/753 (19.8%) 1.918 1.001–3.559 0.038a
5–10 Inhaled bronchodilator 10/38 (26.3%) 95/491 (19.3%) 1.487 0.622–3.290 0.295
Inhaled steroid 7/38 (18.4%) 62/491 (12.6%) 1.561 0.556–3.820 0.316
Asthma-related physician visit 10/38 (26.3%) 82/491 (16.7%) 1.779 0.741–3.955 0.179
Abbreviations: CDH, congenital diaphragmatic hernia; OR, odds ratio.

a
p < 0.05.

Table 2 Results of the secondary analysis between 0 and 5 years of age and 5 and 10 years of age
Age Birth characteristic Bronchodilator Steroid Asthma-related

(years) prescriptions prescriptions physician visits
OR 95% CI p-Value OR 95% CI p-Value OR 95% CI p-Value
a
0–5 Gestational age (wk) 0.66 0.43–0.91 0.024 0.72 0.47–1.00 0.080 0.68 0.45–0.94 0.032a
Birthweight (kg) 0.17 0.03–0.59 0.013a 0.43 0.12–1.28 0.153 0.31 0.08–0.96 0.056
Right-sided defect 1.71 0.38–7.38 0.470 3.08 0.66–22.2 0.187 1.07 0.20–4.65 0.933
Large defect size 0.87 0.20–3.28 0.845 2.36 0.65–9.95 0.207 1.21 0.31–4.42 0.773
Absence of hernia sac 1.29 0.32–6.56 0.737 1.05 0.27–4.02 0.940 0.76 0.19–3.31 0.697
Liver herniated 0.47 0.12–1.62 0.252 1.34 0.43–4.35 0.621 0.71 0.19–2.36 0.586
Pulmonary hypertension 1.88 0.60–6.08 0.284 2.60 0.85–8.52 0.101 1.88 0.60–6.08 0.284
5–10 Gestational age (wk) 0.74 0.31–1.63 0.463 0.76 0.36–1.50 0.436 0.97 0.48–1.92 0.922
Birthweight (kg) 0.65 0.10–4.16 0.646 0.44 0.08–2.16 0.324 0.72 0.14–3.56 0.679
Right-sided defect 3.73 0.41–29.0 0.203 2.08 0.24–14.9 0.463 2.08 0.24–14.9 0.463
Large defect size 0.70 0.03–5.59 0.764 1.68 0.20–10.8 0.594 4.13 0.62–27.4 0.131
Absence of hernia sac 1.83 0.25–37.6 0.605 0.96 0.17–7.52 0.960 0.96 0.17–7.52 0.960
Liver herniated 0.46 0.02–3.31 0.500 0.86 0.11–4.65 0.866 0.31 0.02–2.14 0.308
Pulmonary hypertension 5.25 0.86–43.7 0.084 1.80 0.36–8.71 0.459 4.76 0.93–28.7 0.067
a
p < 0.05.
of respiratory tract infections decreased with age.16 Pre- tions, inhaled steroid prescriptions, or asthma-related physi-
ndergast et al compared lung function in CDH children to cian visits. Between 5 and 10 years of age, we did not find any of
children with anterior wall defects at a median of 11 (range the clinical characteristics to be associated with increased
of 6–24) months and found that CDH was associated with rates of inhaled bronchodilator prescriptions, inhaled steroid
worse long-term lung function.17 Haliburton et al analyzed prescriptions, or asthma-related physician visits.
pulmonary function tests in 33 CDH survivors between 5 and Numerous studies have shown that low birthweight, pre-
17 years of age and found that they had below normal forced maturity, large defect size, right sided defects, liver herniation,
expiratory volume in 1 second (FEV1) and FEV1/FVC (forced absence of a hernia sac, and severe pulmonary hypertension
vital capacity) results.18 Panitch et al analyzed pulmonary have a negative impact on neonatal outcomes.2,3,13,20–29 For
function tests in 98 children with CDH in their first example, the presence of a hernia sac in CDH has been
44 months of age and found that lung function remained associated with decreased time spent on ventilation,30 and
abnormal for this time period.19 Similar to these studies on the presence of severe pulmonary hypertension has been
long-term outcomes of children with CDH, our study sug- associated with mortality.22,31 Despite the previously reported
gested that children with CDH had impaired pulmonary associations with poor neonatal outcomes, we found that only
function between 0 and 5 years of age. However, our study lower birthweights and lower gestational ages were associated
also analyzed respiratory outcomes between 5 and 10 year of with worse respiratory outcomes between 0 and 5 years of age.
age. The cited examples did not follow-up CDH survivors this Furthermore, none of these characteristics were associated
long with the exception of Haliburton et al and El Chehadeh with worse respiratory outcomes between 5 and 10 years of age.
et al. Additionally, to our knowledge, this was the first study This study has the inherent limitations of a retrospective
that was able to compare the long-term outcomes of children cohort study. Additionally, we collected information from a
with CDH to date of birth-matched controls. cohort that was managed over 22 years; advances in treat-
As a secondary objective of this study, we performed ment over time may have created a disparate cohort which
subgroup comparisons of the CDH children to determine if may have impacted our results. Another limitation was that
lower birthweights, lower gestational ages, right-sided 63 of 900 (7.0%) children in the 0 to 5 years of age analysis,
defect, large defect size, liver herniation, absence of a hernia and 77 of 627 (12.3%) children in the 5 to 10 years of age
sac, and severe pulmonary hypertension were associated with analysis were excluded because their data was incomplete. A
increased rates of inhaled bronchodilator prescriptions, third limitation was that the 5 to 10 years of age analyses only
inhaled steroid prescriptions, or asthma-related physician contained 38 children with CDH; this sample size may have
visits. Between 0 and 5 years of age, we found that lower been too small to detect differences in outcomes. A fourth
birthweights and gestational ages were associated with limitation was that we could not control for confounding
increased rates of inhaled steroid prescriptions and asthma- variables. The datasets used to select controls did not contain
related physician visits. No other clinical characteristics were birth characteristics, so we were unable to match birth
associated with increased inhaled bronchodilator prescrip- characteristics of cases and controls, other than age. Despite

these limitations, our study is unique in that we used 6 Danzer E, Gerdes M, D’Agostino JA, et al. Preschool neurological
population-level administrative data. This is a robust, unbi- assessment in congenital diaphragmatic hernia survivors: out-
ased dataset that provided a larger cohort than used in come and perinatal factors associated with neurodevelopmental
impairment. Early Hum Dev 2013;89(06):393–400
previous studies. In addition, we compared the outcomes
7 Safavi A, Synnes AR, O’Brien K, Chiang M, Skarsgard ED, Chiu PPL;
of the children with CDH to date of birth-matched peers, Canadian Pediatric Surgery Network. Multi-institutional follow-
which was not done in previous studies. This enabled us to up of patients with congenital diaphragmatic hernia reveals
quantify differences in the medical outcomes between chil- severe disability and variations in practice. J Pediatr Surg 2012;
dren with CDH and age-matched controls. We hope that this 47(05):836–841
8 Rocha G, Azevedo I, Pinto JC, Guimarães H. Follow-up of the
data will help guide the long-term medical management of
survivors of congenital diaphragmatic hernia. Early Hum Dev
the children with CDH.
2012;88(04):255–258
9 Jancelewicz T, Vu LT, Keller RL, et al. Long-term surgical out-
comes in congenital diaphragmatic hernia: observations from a
Conclusion single institution. J Pediatr Surg 2010;45(01):155–160, discus-
The results of our study suggest that more CDH survivors sion 160
10 Chen C, Jeruss S, Chapman JS, et al. Long-term functional impact of
were prescribed inhaled bronchodilators, inhaled steroids,
congenital diaphragmatic hernia repair on children. J Pediatr Surg
and had asthma-related physician visits in the first 5 years of 2007;42(04):657–665
life compared with age-matched controls. However, between 11 Jutte DP, Roos LL, Brownell MD. Administrative record linkage as a
5 and 10 years of age, these differences were no longer tool for public health research. Annu Rev Public Health 2011;
observed. These findings suggest that CDH survivors have 32:91–108
worse respiratory function compared with their peers, but 12 Roos LL, Jarmasz JS, Martens PJ, et al. Health services informa-
tion: from data to policy impact (25 Years of health services and
only up to 5 years of age. In addition, between 0 and 5 years of
population health research at the Manitoba Centre for Health
age, lower birthweights were associated with and lower Policy). In: Sobolev B, Levy A, Goring S, eds. Data and Measures
gestational ages were associated with increased rates of in Health Services Research. Boston, MA: Springer U.S.; 2016:
inhaled steroid prescriptions and asthma-related physician 1–20
visits and lower gestational ages were associated with higher 13 Morini F, Valfrè L, Capolupo I, Lally KP, Lally PA, Bagolan P;
Congenital Diaphragmatic Hernia Study Group. Congenital dia-
rates of inhaled steroid prescriptions. This association was
phragmatic hernia: defect size correlates with developmental
not found for other clinical characteristics (large defect size,
defect. J Pediatr Surg 2013;48(06):1177–1182
right-sided defect, liver herniation, absence of a hernia sac, or 14 El Chehadeh K, Becmeur F, Weiss L. [Medium and long-term
a severe pulmonary hypertension). Between 5 and 10 years respiratory outcome in patients operated from congenital dia-
of age, no clinical characteristics were found to be associated phragmatic hernia: from a series of 56 patients]. Rev Pneumol
with increased rates of inhaled bronchodilator prescriptions, Clin 2018;74(06):467–482
15 Koziarkiewicz M, Taczalska A, Piaseczna-Piotrowska A. Long-term
inhaled steroid prescriptions, or asthma-related physician
follow-up of children with congenital diaphragmatic hernia—
visits. observations from a single institution. Eur J Pediatr Surg 2014;
24(06):500–507
Funding 16 Kamata S, Usui N, Kamiyama M, et al. Long-term follow-up of
patients with high-risk congenital diaphragmatic hernia. J Pediatr
This study was funded by the Children’s Hospital Research
Surg 2005;40(12):1833–1838
Institute of Manitoba, (Grant/Award Number: OG2017- 17 Prendergast M, Rafferty GF, Milner AD, et al. Lung function at
07) and Thorlakson Chair in Surgical Research. follow-up of infants with surgically correctable anomalies.
Pediatr Pulmonol 2012;47(10):973–978
Conflict of Interest 18 Haliburton B, Mouzaki M, Chiang M, et al. Pulmonary function
and nutritional morbidity in children and adolescents with
None declared.
congenital diaphragmatic hernia. J Pediatr Surg 2017;52(02):
252–256
References 19 Panitch HB, Weiner DJ, Feng R, et al. Lung function over the first
1 The International Center on Birth Defects. International Clearing- 3 years of life in children with congenital diaphragmatic hernia.
house for Birth Defects Surveillance and Research Annual Report Pediatr Pulmonol 2015;50(09):896–907
2014. Rome, Italy: The International Centre on Birth Defects- 20 Delaplain PT, Zhang L, Chen Y, et al. Cannulating the contra-
ICBDSR Centre; 2014:31 indicated: effect of low birth weight on mortality in neonates
2 de Buys Roessingh AS, Dinh-Xuan AT. Congenital diaphragmatic with congenital diaphragmatic hernia on extracorporeal mem-
hernia: current status and review of the literature. Eur J Pediatr brane oxygenation. J Pediatr Surg 2017;52(12):2018–2025
2009;168(04):393–406 21 Jaillard SM, Pierrat V, Dubois A, et al. Outcome at 2 years of infants
3 Skari H, Bjornland K, Haugen G, Egeland T, Emblem R. Congenital with congenital diaphragmatic hernia: a population-based study.
diaphragmatic hernia: a meta-analysis of mortality factors. J Pediatr Ann Thorac Surg 2003;75(01):250–256
Surg 2000;35(08):1187–1197 22 Brindle ME, Cook EF, Tibboel D, Lally PA, Lally KP; Congenital
4 Peetsold MG, Heij HA, Kneepkens CMF, Nagelkerke AF, Huisman J, Diaphragmatic Hernia Study Group. A clinical prediction rule for
Gemke RJBJ. The long-term follow-up of patients with a congeni- the severity of congenital diaphragmatic hernias in newborns.
tal diaphragmatic hernia: a broad spectrum of morbidity. Pediatr Pediatrics 2014;134(02):e413–e419
Surg Int 2009;25(01):1–17 23 Putnam LR, Harting MT, Tsao K, et al; Congenital Diaphragmatic
5 Takayasu H, Masumoto K, Jimbo T, et al. Analysis of risk factors of Hernia Study Group. Congenital diaphragmatic hernia: defect size
long-term complications in congenital diaphragmatic hernia: a and infant morbidity at discharge. Pediatrics 2016;138(05):
single institution’s experience. Asian J Surg 2017;40(01):1–5 e2016–e2043

24 Spaggiari E, Stirnemann J, Bernard JP, De Saint Blanquat L, 28 Colvin J, Bower C, Dickinson JE, Sokol J. Outcomes of congenital
Beaudoin S, Ville Y. Prognostic value of a hernia sac in congenital diaphragmatic hernia: a population-based study in Western
diaphragmatic hernia. Ultrasound Obstet Gynecol 2013;41(03): Australia. Pediatrics 2005;116(03):e356–e363
286–290 29 Grizelj R, Bojanić K, Vuković J, et al. Hernia Sac presence portends
25 Zamora IJ, Cass DL, Lee TC, et al. The presence of a hernia sac in better survivability of isolated congenital diaphragmatic hernia
congenital diaphragmatic hernia is associated with better fetal lung with “liver-up”. Am J Perinatol 2017;34(05):515–519
growth and outcomes. J Pediatr Surg 2013;48(06):1165–1171 30 Levesque M, Derraugh G, Schantz D, et al. The presence of a hernia
26 Panda SS, Bajpai M, Srinivas M. Presence of hernia sac in prediction sac in isolated congenital diaphragmatic hernia is associated with
of postoperative outcome in congenital diaphragmatic hernia. less disease severity: a retrospective cohort study. J Pediatr Surg
Indian Pediatr 2013;50(11):1041–1043 2019;54(05):899–902
27 Bouchghoul H, Marty O, Fouquet V, et al. Congenital diaphrag- 31 Coughlin MA, Werner NL, Gajarski R, et al. Prenatally diagnosed
matic hernia has a better prognosis when associated with a hernia severe CDH: mortality and morbidity remain high. J Pediatr Surg
sac. Prenat Diagn 2018;38(09):638–644 2016;51(07):1091–1095

Original Article
Ultrasound Monitoring after Pelvis-Sparing

Dismembered Pyeloplasty: High Sensitivity and
Low Specificity for the Success of Operation
Rim Kiblawi1 Joachim F. Kuebler1 Claus Petersen1 Benno M. Ure1 Alejandro D. Hofmann1
1 Department of Pediatric Surgery, Hannover Medical School, Address for correspondence Rim Kiblawi, MD, Department of
Hannover, Germany Pediatric Surgery, Hannover Medical School, Carl-Neuberg-Str.1,
Hannover 30625, Germany (e-mail: [email protected]).
Eur J Pediatr Surg
Abstract Aim A decrease in the anteroposterior diameter (APD) of the renal pelvis on
ultrasound has been postulated to be indicative of sufficient pelvic drainage after
pyeloplasty. Traditionally, pyeloplasty is combined with a reduction of the renal pelvis.
We have recently demonstrated that resection of the pelvis during pyeloplasty is not
necessary. We aimed to evaluate the efficacy of ultrasound APD measurements during
follow-up to identify sufficient pelvic drainage in these patients.
Materials and Methods Data from children (0–16 years) who underwent pelvis-
sparing pyeloplasty in our institution from 2007 to 2018 were analyzed retrospectively.
We included only those patients for whom pre- and postoperative ultrasound and renal
scan data were available. Patients with a decrease versus patients with an increase in
APD were analyzed with regard to urinary drainage and reoperation.
Results Seventy-three patients who underwent follow-up at a mean of 3 months after
operation were included; 61 showed a decrease in APD. Renal scan showed sufficient
urinary drainage in 58 of them, with none requiring reoperation. Twelve patients had an
increase in APD. Six of these showed free urinary drainage on renal scan; another six showed
insufficient drainage, of whom five required reoperation. The positive predictive value of a
Keywords decrease in APD was 1, and the negative predictive value of increase in APD was 0.42.
► pyeloplasty Conclusion To our knowledge, this is the first study evaluating the efficacy of
► hydronephrosis ultrasound measurements to identify patients with decompensated urinary drainage
► anteroposterior during early follow-up after pyeloplasty with pelvis sparing. Post- versus preoperative
diameter decrease in renal pelvis diameter appears to be sufficient to rule out recurrence of
► ultrasound obstruction. Renal scan seems to be indicated only in cases with post- versus
monitoring preoperative increase in the APD of the renal pelvis on ultrasound.
Introduction Success after pyeloplasty is defined by radiographic and

clinical criteria. Today, MAG3 renal scintigraphy is commonly
Ureteropelvic junction obstruction (UPJO) represents a used to timely detect recurrent obstruction after pyeloplasty,
common cause of unilateral hydronephrosis in children.1 which can lead to loss of renal function.2
Dismembered pyeloplasty (open or laparoscopically) is the However, to date, there are no commonly approved
gold standard therapy for UPJO, with high success rates recommendations to guide the modality of image surveil-
reported in children.2,3 Goals of the operative correction lance, timing, or duration of follow-up after pyeloplasty.2
are to preserve renal function and to minimize symptoms Furthermore, with growing awareness of radiation, in par-
such as pain and infection, if present. ticular in the pediatric population, several authors have

accepted ISSN 0939-7248.
June 15, 2019
US Monitoring after Pelvis-Sparing Pyeloplasty Kiblawi et al.
recently advocated that ultrasound could replace renal scan or MAG3. Urinary drainage was calculated as the radionu-
in the follow-up of children after pyeloplasty.2,4,5 A mea- clide excretion, time to reach the maximum of the curve or
surement that has proven to be of most value is the ante- half clearance time [T/2] >20 minutes; after diuretic admin-
roposterior diameter (APD) of the renal pelvis.6,7 The APD istration. Sufficient urinary drainage (>50%) was considered
was postulated to be indicative of sufficient pelvic drainage when the renal washout occurs in 30 minutes after diuretic
during early follow-up.4,5 administration. When washout exceeded 30 minutes after
A cause of potential bias, however, is the fact that pyelo- diuretic injection, the drainage was judged to be insufficient
plasty nowadays is performed with varying degrees of renal (< 50%), and further diagnostics or operation was required.
pelvis reduction. This may impact the interpretation of the The primary study outcome was to measure the efficacy of
postoperative result. We have previously demonstrated that D-APD to predict the success of pyeloplasty. Success was
reduction of the dilated renal pelvis during pyeloplasty in defined as the absence of recurrence of UPJO. Criteria to define
children is not necessary since there were no differences in recurrence of UPJO were urinary drainage less than 50%, a
the renographic outcome of patients treated with and with- decrease in RKU of more than 10%, increase in hydronephrosis
out reduction of the pelvis.8 and APD, pyelonephritis despite antibiotic prophylaxis, and
In this study, the potential role of ultrasound to replace abdominal pain.9
renal scan during follow-up in children who underwent The secondary study outcome was the correlation between
pyeloplasty with pelvis sparing was investigated. The aim D-APD (ultrasound) and urinary drainage (renal scan).
of the study was to determine whether postoperative sono-
graphic APD measurement could be used as a sentinel Statistical Methods
follow-up screening method in these patients to select those All numerical data are presented as mean standard error of
who may need further diagnostics such as renal scan. the mean or as median (range). Student’s t-test was used for
the evaluation of differences between two normal distributed
groups. The confidence interval was set at 95%.
Statistical analysis was calculated with Fisher’s test using
We conducted a comprehensive retrospective chart review of SPSS statistical software (SPSS GmbH, Munich, Germany) for
all the patients (0–16 years) who underwent pyeloplasty with the correlation between APD and the binary prediction
pelvis sparing in our institution from 2007 to 2018 following model for the outcome of interest (successful pyeloplasty:
ethics board approval. Inclusion criteria were laparoscopic or yes/no). Statistical significance was given when p was <0.05.
open pyeloplasty and the availability of both renal scan and The linear regression curve between continuous variables
ultrasound pre- and postoperatively at follow-up after (variation of urinary drainage vs. variation in APD) was calcu-
3 months. Symptoms, indication for intervention, age at lated using Prism 8 statistical software (Version 8.1.1, Graph-
operation, side, and results of pre- and postoperative imaging Pad). Results are presented as mean standard error of the
studies were assessed. mean. Statistical significance was given when p was <0.05.
Pyeloplasty was performed in a dismembered fashion
without reduction of the dilated renal pelvis. All the children
Results
were stented with either an internal (for 6 weeks) or external
(for 1 week) nephroureteral catheter. A total of 147 pyeloplasties were performed from January 2007
The results of APD measurements at renal ultrasound and to November 2018 in our institution. Of 147 pyeloplasties, 73
the drainage rates at renal scan were analyzed at the first (49.7%) met all inclusion criteria. In 66 of 73 patients, the
postoperative follow-up, which was usually scheduled at indication for surgery was confirmed by urinary drainage
3 months. APD on ultrasound was defined as the transverse < 50%. In seven patients, the indication for surgery was not
diameter bordered by renal parenchyma edges. The measure- confirmed by insufficient urinary drainage. Five of these seven
ments were performed by a pediatric radiologist or pediatric patients presented with symptoms such as recurrent pyelone-
surgeon of the urology team. In each patient, APD was measured phritis or abdominal pain. Two patients were newborns
at least twice, pre- and postoperatively at 3 months after stent with prenatal diagnosis and postnatally worsening of
removal. Difference in APD was determined as D-APD ¼ pre- hydronephrosis.
operative APD – postoperative APD. Patients underwent pyeloplasty at a median age of
Positive value denoted a decrease in the pelvis diameter 19 months (range: 1–170 months) (mean: 48.3 months;
and thereby an improvement. Negative value denoted an standard deviation 51.7). Of 73 pyeloplasties, 65 (89%)
increase in the pelvis diameter and thereby a potential were performed minimally invasively, of which 3 were con-
worsening of the urinary drainage. verted to open surgery (►Table 1).
Two groups were considered: group A included patients Postoperative monitoring included ultrasound at a mean
who experienced decrease in the APD, and group B included of 130 (SD 107) postoperative days (median: 99 days;
patients who experienced increase in the APD at 3-month range: 12–432) and renal scan at a mean of 105 (SD 72.9)
follow-up. postoperative days (median: 100; range: 8–553) (p ¼ 0.06).
In renal scan, documented parameters were relative The average preoperative APD on ultrasound in group A
kidney uptake (RKU) and urinary drainage expressed in was 32 mm (SD 14.8 mm) and the average postoperative
percentage and performed with 123J-orthoiodohippurate APD was 16.7 mm (SD 10.5 mm); in group B, the average

Table 1 Patients characteristics
Total Decrease in APD Increase in APD p–Value

(decrease vs. Increase)
Number of pyeloplasties (n) 73 61 12 1
Median age at operation, 19 (1–170; 19 (1–171); 19(1–147); 0.89
months (range); mean SD) 48.3 51.7) 48.7 51.2 48.7 51.6
Surgical technique 0.68
Laparoscopic (n) 62 51 11
Open retroperitoneal (n) 11 10 1
Side 0.53
Right (n) 30 24 6
Left (n) 43 37 6
Median time to first follow-up,
days (range); mean, SD
Renal scan 100 (8– 553); 98 (8–553); 94 (58–113; 0.43
105 72.9 108 78.6 90 17.6)
Ultrasound 99 (12–432); 101(12–432); 91 (46–297); 0.46
130 107 136 113 112 70.1
was 24 mm (SD 13.7 mm) preoperatively versus 31 mm In group A (decrease in APD; n ¼ 61), the renal scan
(SD 18.6 mm) postoperatively. confirmed sufficient drainage in 58 of 61 patients. In 3 of
Group A versus group B did not exhibit any statistically these 61 patients, the scan showed no sufficient drainage of
significant difference regarding age, surgical techniques, and the kidney despite a decrease in the APD of the renal pelvis
other parameters (p > 0.05) (►Table 1), except for the inci- on ultrasound. These three patients suffered from other
dence of recurrence (p ¼ 0.001). comorbidities such as primary obstructive megaureter and
In group A, on renal scan, the mean preoperative urinary vesicoureteral reflux. None of these patients needed reoper-
drainage was 30% (SD 17%) with an RUK of 44% (SD ation during close follow-up. Therefore, all the patients in
18%), and the mean postoperative urinary drainage was group A (61/61) had a successful pyeloplasty at 3-month
70% (SD 14%) with an RUK of 45 (SD 15%) on the affected follow-up .
side. In group B, the mean urinary drainage was 31% (SD The positive predictive value of a decrease in APD in
18%) with a mean RUK of 43% (SD 10%) preoperatively, predicting success was 1, the sensitivity for success was
and 59% (SD 31%) with a mean RUK of 34% (SD 15%) 100%, and the specificity was 89% (►Fig. 2).
postoperatively. In group B (increase in APD; n ¼ 12), six of 12 patients
Pre-and postoperative variation in the difference of APD experienced sufficient urinary drainage on renal scan and
in ultrasound as well as the urinary drainage at renal scan therefore were considered having a successful pyeloplasty
showed a highly significant direct correlation in the linear despite an increase in APD. Renal scan confirmed insufficient
regression curve (p ¼ 0.0004) (►Fig. 1). drainage in six patients. One of these patients additionally
suffered from distinct vesicoureteral reflux without signs of
obstruction at the ureteropelvic junction. Five of 12 patients
required a redo pyeloplasty. Therefore, the negative predictive
value was 0.42. All recurrences had an increase in APD on
ultrasound (sensitivity of 42% and specificity of 100%) (►Fig. 2).
In four of the patients who suffered from recurrent UPJO,
scan was performed at an average of 90.2 days (SD 11.2),
whereas ultrasound was performed at 89 days (SD 12.7)
after pyeloplasty (►Table 2). No statistically significant
difference was found between renal scan and ultrasound
regarding the time interval between initial operation and
first follow-up (p ¼ 0.31). All recurrences of obstruction
were detected at 3-month follow-up, and all of them
presented with both increasing APD on ultrasound and
Fig. 1 Correlation of differences (%) between preoperative and
postoperative urinary drainage (renal scan) as well as APD (ultra-
insufficient urinary drainage on renal scan. Two redo pyelo-
sound); (p ¼ 0.0004). APD, anteroposterior diameter of the renal plasties were performed laparoscopically and three conven-
pelvis measured on ultrasound. tionally through a retroperitoneal approach.

Fig. 2 Follow-up results in 73 patients who underwent pre- and postoperative ultrasound measurements of the diameter of the renal pelvis
(APD) and renal scan. APD, anteroposterior diameter of the renal pelvis measured on ultrasound. Sufficient urinary drainage (>50%) ¼ renal
washout occurs in 30 minutes after diuretic administration measured on renal scan.
Table 2 Patients with recurrence: characteristics
Patient 1 Patient 2 Patient 3 Patient 4 Patient 5

Age at initial pyeloplasty (months) 99 91 7 2 1
Surgical technique LAP LAP LAP Open LAP
Postoperative days of ultrasound 96 69 91 88 103
D-APD (mm) –45 –5 –5 –7 –3
Postoperative days of renal scan 96 72 92 89 102
Urinary drainage postoperative 23% 40% 12% 34% 30%
Time to redo pyeloplasty (days) 117 86 94 101 112
Abbreviations: D-APD, difference of anteroposterior diameter; LAP, laparoscopic pyeloplasty.

Note: “Open” indicates conventional retropertitoneal pyeloplasty.
Discussion value.6,7 Romao et al postulated that the change in the

APD may indicate which patients require closer follow-up
Although success rates after dismembered pyeloplasty are monitoring and may help identify patients at risk for
high, there is no general consensus on how and when reintervention.5 Similarly, Rickard et al demonstrated in
postoperative follow-up should be performed. Although 138 patients that a reduction in APD of more than 40% was a
consistent guidelines on the modality of image surveillance useful indicator of successful pyeloplasty.4 However, in both
and timing after pyeloplasty2 are lacking, renal scan and studies, the renal pelvis had been routinely resected during
renal ultrasound have evolved as the diagnostic corner- pyeloplasty.
stones.10 However, the use of routine and repeated renal All children of our series underwent pelvis-sparing pye-
scans has been challenged during the last decade due to loplasty, as we previously demonstrated that reduction of
radiation exposure.11,12 It has been suggested that ultra- the pelvis in laparoscopic dismembered pyeloplasty is
sound could be the sole initial method to be used at follow-up unnecessary.8 Thus, the size of the pelvis on our patients
and that renal scan may be reserved for selected cases.4,6,13 was not altered by the operation, which may allow a more
Several sonographic parameters for the evaluation of valid comparison of the pre- and postoperative APD values.
obstruction after pyeloplasty have been reviewed.6 The An improvement in the APD at the follow-up was identified
grade of hydronephrosis at follow-up is a morphological in 83.5% of our patients. Three of these patients showed
parameter that may not identify relevant obstruction.4 insufficient urinary drainage at renal scan, but that was due
Consequentially, several authors focused on an objective to comorbidities such as vesicoureteral reflux and primary
and quantifiable sonographic measure for relevant obstruc- obstructive megaureter. None of these patients required
tion.4,5,14 In several studies, the difference of APD post- reoperation. This indicates that a decrease in the APD indi-
versus preoperative has been suggested to be of most cates success of pyeloplasty, with a sensitivity of 100%.

On the other hand, 12 (16.5%) of our patients experienced a preoperative decrease in APD of renal pelvis on ultrasound
postoperative increase in APD, but only in five (6.8%) patients a appears to be sufficient to rule out recurrence of obstruction.
redo pyeloplasty for significant obstruction was necessary. Renal scan seems to be indicated only in cases with a post-
Thus, an APD increase did not correlate with the recurrence of versus preoperative increase in the APD of renal pelvis on
obstruction as its negative predictive value was only 42%. ultrasound. Thus, our study provides further support to limit
These results are in line with two previous reports showing postoperative tests and avoid radiation exposure in most
that only 20 to 28% of the patients with an increase in hydro- children after pyeloplasty.
nephrosis at 3-month follow-up required a redo pyelo-
plasty.15,16 The authors of these studies concluded that a Conflict of Interest
favorable or indeterminate ultrasound had a high negative None declared.
predictive value for recurrent obstruction. Both studies, how-
ever, based their results on grades of hydronephrosis as
defined by the Society of Fetal Urology. More recent studies References
defend this approach but also advocate the use of more precise 1 Aksu N, Yavaşcan O, Kangin M, et al. Postnatal management of
infants with antenatally detected hydronephrosis. Pediatr Neph-
parameters such as the APD, as used in our study.5
rol 2005;20(09):1253–1259
Renal scan is still the gold standard to evaluate sufficient 2 Hsi RS, Holt SK, Gore JL, Lendvay TS, Harper JD. National trends in
renal pelvic drainage and is widely performed after followup imaging after pyeloplasty in children in the United
pyeloplasty.11,17 The main benefit of nuclear studies is States. J Urol 2015;194(03):777–782
the validated assessment of renal function and confirma- 3 Radmayr C, Bogaert G, Dogan HS, et al. EAU Guidelines on
tion of adequate urinary drainage.10,11 Nonetheless, Paediatric Urology. Available at: https://uroweb.org/guideline/
paediatric-urology/. Accessed July 9, 2019
nuclear scanning requires intravenous access, sometimes
4 Rickard M, Braga LH, Oliveria J-P, Romao R, Demaria J, Lorenzo AJ.
urethral catheterization, and entails radiation exposure as Percent improvement in renal pelvis antero-posterior diameter
well as incurring additional cost.2 Furthermore, the utility (PI-APD): prospective validation and further exploration of cut-
of diuretic renal scans in reliably diagnosing obstruction in off values that predict success after pediatric pyeloplasty sup-
children has been previously questioned both in the porting safe monitoring with ultrasound alone. J Pediatr Urol
2016;12(04):228.e1–228.e6
pre-and postoperative management of UPJO.15,18,19 Nine
5 Romao RLP, Farhat WA, Pippi Salle JL, et al. Early postoperative
patients of our cohort did not drain sufficiently on renal
ultrasound after open pyeloplasty in children with prenatal
scan. However, only five needed redo pyeloplasty. The hydronephrosis helps identify low risk of recurrent obstruction.
remaining four patients showed other obstructive patterns J Urol 2012;188(06):2347–2353
on renal scan due to comorbidities after adjustment of the 6 Fernández-Ibieta M, Nortes-Cano L, Guirao-Piñera MJ, Zambudio-
region of interest of urinary drainage, such as primary Carmona G, Ruiz-Jiménez JI. Radiation-free monitoring in the
long-term follow-up of pyeloplasty: are ultrasound new param-
obstructive megaureter and vesicoureteral reflux. In these
eters good enough to evaluate a successful procedure? J Pediatr
patients, no further surgical intervention was necessary. Urol 2016;12(04):230.e1–230.e7
We determined a highly significant direct correlation 7 Bouzada MCF, Oliveira EA, Pereira AK, et al. Diagnostic accuracy of
between the pre- and postoperative variation of APD on fetal renal pelvis anteroposterior diameter as a predictor of
ultrasound versus urinary drainage on renal scan. Given the uropathy: a prospective study. Ultrasound Obstet Gynecol
high positive predictive value of a decrease in APD for the 2004;24(07):745–749
8 Reismann M, Gratz KF, Metzelder M, Ure B, Glüer S. Excision of the
success of pyeloplasty, we suggest that the difference in pre-
dilated pelvis is not necessary in laparoscopic dismembered
and postoperative APD could be used as a sentinel screening pyeloplasty. Eur J Pediatr Surg 2008;18(01):19–21
method to select patients who need further diagnostics. 9 Metzelder ML, Schier F, Petersen C, Truss M, Ure BM. Laparoscopic
Our study has some limitations. The validity of our findings transabdominal pyeloplasty in children is feasible irrespective of
is limited by the retrospective nature of the study, the small age. J Urol 2006;175(02):688–691
number of patients, in particular those with recurrence of 10 Thomas JC, DeMarco RT, Donohoe JM, Adams MC, Pope JC IV,
Brock JW III. Management of the failed pyeloplasty: a contempo-
pelvic obstruction, and a short follow-up period of a mean of
rary review. J Urol 2005;174(06):2363–2366
3 months. Hence, more data and in particular longer follow-up 11 Pohl HG, Rushton HG, Park JS, Belman AB, Majd M. Early diuresis
are required before a final recommendation can be given. In renogram findings predict success following pyeloplasty. J Urol
addition, it has to be pointed out that the results of this study 2001;165(6 Pt 2):2311–2315
only account for patients who undergo pelvis-sparing pyelo- 12 van den Hoek J, de Jong A, Scheepe J, van der Toorn F, Wolf-
fenbuttel K. Prolonged follow-up after paediatric pyeloplasty: are
plasty. Furthermore, the evaluation of the APD may depend on
repeat scans necessary? BJU Int 2007;100(05):1150–1152
the investigator and state of hydration. 13 Babu R, Sai V. Pelvis/cortex ratio: a sonographic marker of pelvi
ureteric junction obstruction in children. Indian J Urol 2010;26
(04):494–496
Conclusion 14 Babu R, Sai V. Pelvis/cortex ratio: an early marker of success follow-
To our knowledge, this is the first study evaluating the ing pyeloplasty in children. J Pediatr Urol 2010;6(05):473–476
15 Cost NG, Prieto JC, Wilcox DT. Screening ultrasound in follow-up
efficacy of ultrasound measurements of the diameter of
after pediatric pyeloplasty. Urology 2010;76(01):175–179
the renal pelvis to identify potential patients with decom- 16 Almodhen F, Jednak R, Capolicchio J-P, Eassa W, Brzezinski A, El-
pensated urinary drainage during early follow-up after pye- Sherbiny M. Is routine renography required after pyeloplasty? J Urol
loplasty without reduction of the renal pelvis. A post- versus 2010;184(03):1128–1133

17 Burgu B, Suer E, Aydogdu O, Soygur T. Pelvic reduction during sign of obstruction in children with a prenatal diagnosis of
pyeloplasty for antenatal hydronephrosis: does it affect outcome unilateral renal pelvic dilatation. J Urol 2003;169(05):1828–1831
in ultrasound and nuclear scan postoperatively? Urology 2010;76 19 Wang T-M, Chang P-L, Kao P-F, Hsieh M-L, Huang S-T, Tsui K-H.
(01):169–174 The role of diuretic renography in the evaluation of obstructed
18 Amarante J, Anderson PJ, Gordon I. Impaired drainage on diuretic hydronephrosis after pediatric pyeloplasty. Chang Gung Med J
renography using half-time or pelvic excretion efficiency is not a 2004;27(05):344–350

Original Article
Liver Organoids Generated from Mice with Necrotizing

Enterocolitis Have Reduced Regenerative Capacity
Hiromu Miyake1,2 Carol Lee1 Shogo Seo1 Bo Li1 Agostino Pierro1
1 Division of General and Thoracic Surgery, Hospital for Sick Children, Address for correspondence Agostino Pierro, MD, Hospital for Sick
Toronto, Ontario, Canada Children, 555 University Avenue, Suite 1526, Toronto, ON, M5G 1X8,
2 Department of Pediatric Surgery, Shizuoka Children’s Hospital, Canada (e-mail: [email protected]).
Shizuoka, Japan
Eur J Pediatr Surg
Abstract Introduction Necrotizing enterocolitis (NEC) is one of the most severe gastrointesti-
nal diseases in infancy. NEC can cause metabolic derangements, multi-organ injury
including severe liver damage. The mechanism leading to hepatic damage in NEC
remains unclear. The aim of this study is to establish and characterize liver organoids
from NEC mice.
Materials and Methods Following ethical approval (#44032), we induced experimen-
tal NEC from postnatal day 5 (P5) to P9 using C57BL/6 mice pups. NEC was induced by
gavage formula feeding, gavage lipopolysaccharide (LPS) administration, and hypoxia.
Breastfed pups were used as control. On P9, NEC and control pups were sacrificed and
liver tissue was harvested and organoids were generated. Organoid size was recorded
daily (day 2–4) by measuring their surface area and organoid growth was calculated.
RNA was extracted on day 4 after liver organoid generation.
Results Organoid growth rate was significantly lower in NEC liver organoids com-
pared to control liver organoids. mRNA expression of liver progenitor cells markers of
LGR5 and SOX9 was lower in NEC liver organoids compared to control liver organoids.
Keywords Similarly, expression of proliferation markers of Ki67 and PCNA was lower in NEC liver
► necrotizing organoids.
enterocolitis Conclusion We report a novel technique to generate liver organoids during NEC.
► liver injury These organoids are characterized by reduced progenitor cells, reduced proliferation,
► organoids and overall impaired regenerative capacity. Liver progenitor cells are important targets
► regeneration to prevent liver damage in NEC and promote recovery.
Introduction multi organ failure including liver, and such patients have
worsened clinical outcomes.4,5 Therefore, it is important to
Necrotizing enterocolitis (NEC) is one of the most severe understand the mechanism of multiorgan damage in NEC and
gastrointestinal diseases in premature infants. Despite numer- to investigate new therapeutic strategies.
ous efforts to improve clinical outcomes of NEC, the incidence It has been reported that NEC is associated with liver injury4
of NEC remains high and mortality and morbidity are high.1–3 and that gut–liver inflammatory interaction is involved in the
It has been reported that sometimes NEC is associated with mechanism of liver injury in NEC.6,7 The liver regenerative
capacity is due to the turnover of mature hepatocytes without a
Bo Li's ORCID is https://orcid.org/0000-0002-8298-8344. direct involvement of stem/progenitor cells.8–10 However,
Agostino Pierro's ORCID is https://orcid.org/0000-0002-6742-6570. when this regenerative capacity is overwhelmed, hepatic

June 15, 2019
Liver Organoids Generated from NEC Mice Miyake et al.
progenitor cells (HPCs) are recruited to repair the damage. The United States). Hepatic duct fragments were then resus-
leucine-rich repeat-containing G-protein coupled receptor 5 pended in matrigel (Corning, Corning, New York, United
(LGR5) is the stem cell marker in the Wnt-driven self-renewing States) and transferred into 24-well plates. After polymeri-
tissues. LGR5 is one of the liver progenitor cell markers involved zation, mouse HepatiCult organoid growth medium (Stem-
in generating functional hepatocytes and cholangiocytes.11 We Cell Technologies, Cambridge, Massachusetts, United States)
reported that expression of progenitor cell marker LGR5 and was overlaid on the gel in each well. Organoids were main-
proliferation markers were impaired in the liver during exper- tained in a 37°C incubator with the culture medium. Surface
imental NEC in pup mice.12 These results suggest that the area of organoids was measured every 24 hours from day 2 to
hepatic regeneration capacity is impaired in NEC. Promoting 4 and then RNA was extracted. To estimate the growth rate of
liver regenerative capacity via stimulation of progenitor cells organoids, their surface area at day 3 and 4 was compared to
may protect the liver from the injury caused by NEC. that at baseline (day 2).
The recent development of 3D cell culture conditions has
allowed the in vitro expansion of primary tissues from healthy RT-qPCR
tissue, as well as from tissue affected by disease. These 3D The expression of messenger RNA (mRNA) was examined by
structures are known as organoids.11,13–15 It has been reported quantitative reverse transcription (RT-qPCR). The mRNA was
that organoids originated from a diseased organ recapitulate extracted from samples by TRIzol Reagent (Thermo Fisher
specific features of the disease itself.13,15 Therefore, organoids Scientific, Inc., Illinois, United States). Complimentary DNA
are used as disease model to understand the mechanism of the (cDNA) was made with qScript cDNA SuperMix (Quantabio,
disease and to test new therapeutic strategies. Beverly, Massachusetts, United States) and S1000 Thermal
The aim of this study was to investigate whether organo- Cycler (Bio-Rad Laboratories, Inc., Hercules, California, Unit-
ids originated from the liver of NEC pup mice have specific ed States). Real-time PCR was performed with advanced
regenerative capacity. qPCR Master Mix and CFX384 Real-Time System (Bio-Rad
Laboratories, Inc., California, United States). The expression
was normalized to the expression of the housekeeping genes
of GAPDH and RPLO. Sequence of primer for each gene is
Experimental Model shown below:
Following ethical approval (#44032), we induced experimen- IL6: sense 5′-CAACTCCATCTGCCCTTCA-3′ and antisense
tal NEC in C57BL/6 mice pups. On postnatal day 5, pups were 5′-TTGTGGGTGGTATCCTCTGT-3′.
randomly assigned to either of control (n ¼ 5) and NEC LGR5: sense 5′-CGAGCCTTACAGAGCCTGATACC-3′ and an-
(n ¼ 5). tisense 5′-TTGCCGTCGTCTTTATTCCATTGG-3′.
NEC was induced by giving hypoxia (10 minutes, four times SOX9: sense 5′- CCACGGAACAGACTCACATCTCTC-3′ and
a day), gavage administration of lipopolysaccharide (LPS; antisense 5′- CTGCTCAGTTCACCGATGTCCACG-3′.
4 mg/kg/day), and gavage formula feeding (15 g Similac Lower Ki67: sense 5′-CAACTTTGGTGATTCCATTA-3′ and anti-
Iron [Abbott Laboratories, Ltd, Saint-Laurent, QC, Canada] sense 5′-TTAGGAGGCAAGTTTTCATC-3′.
þ75 mL Esbilac Puppy Milk Replacer [PetAg, Inc., Hampshire, PCNA: sense 5′-GCCGAGATCTCAGCCATATT-3′ and anti-
Illinois, United States; 152 kcal/100 mL]: 50 μL/g of body sense 5′-ATGTACTTAGAGGTACAAAT-3′.
weight four times a day) from postnatal day 5 to 9 using an GAPDH: sense 5′-GACGAGGCCCAGAGCAAGAGA-3′ and
established model.16,17 Control pups were kept with their antisense 5′-GGGTGTTGAAGGTCTCAAACA-3′.
mother and were breastfed. On postnatal day 9, the pups RPLO: sense 5′- GGCGACCTGGAAGTCCAACT-3′ and anti-
were sacrificed, and the distal ileum and liver were harvested. sense 5′- CCATCAGCACCACAGCCTTC-3′.
HE (Hematoxylin and Eosin) Staining Statistics

Ileum samples were fixed in 4% formalin and embedded in GraphPad Prism 6 (GraphPad Software Inc., San Diego,
paraffin, then stained with hematoxylin and eosin (HE). The California, United States) was used for statistical analyses.
histology slides were blindly evaluated by three independent Continuous data were analyzed using t- test. p-values < 0.05
investigators (B.L., C.L., and S.S.) using a published scoring were considered significant. Data are quoted as mean and SD.
system for NEC severity.17 Mucosal injury score of 2 or more
was consistent with the development of NEC.
Results
Liver Organoid Development NEC Induction and Liver Injury
Liver organoids were cultured according to the protocol from To confirm that NEC was successfully induced, HE staining
StemCell Technologies (Cambridge, Massachusetts, United was performed on postnatal day 9 (P9) in the terminal ileum
States). Liver was harvested from NEC and control pups on of control and NEC pups. NEC mice had significantly higher
postnatal day 9. Hepatic duct fragments were isolated by mucosal injury compared to control (►Fig. 1A and B). All NEC
digestion using tissue dissociation cocktail which includes pups had injury score of 2 (confirmed NEC) consisting in
collagenase IV, dispase, and Dulbecco’s Modified Eagle’s disarrangement of villus enterocytes and severe villus core
medium/nutrient mixture F-12(DMEM/F-12) with 15mM separation. This result indicates that intestinal mucosal
HEPES (StemCell Technologies, Cambridge, Massachusetts, injury was successfully induced in NEC group.

Fig. 1 Validation of NEC induction. (A) HE staining of control and NEC ileum. (B) Mucosal injury score. Injury score was significantly higher in NEC
compared to control (p < 0.05). NEC, necrotizing enterocolitis; HE, hematoxylin and eosin.
Liver injury was tested using qPCR. Liver mRNA was lower in NEC liver organoids compared to control liver organo-
extracted from both control and NEC pups after NEC induction ids (►Fig. 3B). This indicates that NEC liver organoids have less
(P9). Liver inflammation was evaluated by the mRNA expres- growth ability compared to control liver organoids.
sion of inflammatory cytokine of IL6. IL6 mRNA expression was RNA was extracted from organoids at day 4. The mRNA
significantly higher in NEC liver compared to control liver expression of liver progenitor markers LGR5 and SOX9 were
(►Fig. 2A). Similarly, the expression of liver progenitor cell significantly lower in NEC liver organoids compared to
marker LGR5 was lower in NEC liver compared to control liver control liver organoids (►Fig. 4A). Similarly, mRNA expres-
(►Fig. 2B). These results indicate that NEC liver had inflamma- sion of proliferation markers Ki67 and PCNA were signifi-
tion and impaired liver progenitor cell marker expression. cantly lower in NEC liver organoids compared to control liver
organoids (►Fig. 4B).
Organoid Growth and Regenerative Markers
Expression
Discussion
Organoids were derived from liver taken from NEC and control.
Organoid growth was observed daily from day 2 to 4 (48 hours, A novel method has been established for culturing and
72 hours, and 96 hours after culture initiation; ►Fig. 3A). The growing liver organoids from normal mice pups as well as
growth rate (surface area) at day 3 and 4 was significantly from pups exposed to NEC induction. The growth of liver
Fig. 2 Expression of mRNA in liver tissue. (A) mRNA expression of inflammatory cytokine IL6. Expression of IL6 was significantly higher in NEC
compared to control (p < 0.05). (B) mRNA expression of liver progenitor cells LGR5. Expression of LGR5 was significantly lower in NEC compared
to control (p < 0.05). LGR5, leucine rich repeat containing G protein-coupled receptor 5; mRNA, messenger ribonucleic acid.

Fig. 3 Organoids growth. (A) Optical microscope images of organoids from day 2 to 4. (B) Organoids growth rate measured by surface area.
Surface area of each organoids was divided by that of day 2. Organoid growth rate at day 3 and day 4 was significantly lower in NEC compared to
control (day 3: p < 0.05; day 4: p < 0.05). NEC, necrotizing enterocolitis.
organoids was impaired in NEC compared to control. In the liver itself produces inflammatory cytokines such as TNF-α
addition, mRNA expression of liver progenitor markers and IL-18. These cytokines can cause liver injury and can
(LGR5 and SOX9) and proliferation markers (Ki67 and circulate back to the intestine and further aggravate intestinal
PCNA) were significantly lower in NEC liver organoids inflammation and damage.
compared to control liver organoids. These results indicate The liver is known to have good self-renewing capacity.8–10
that liver organoids originated from NEC pups have less Turnover of normal liver is usually achieved by de novo
regenerative capacity. replication of its mature cells, such as hepatocytes and ductal
Severe intestinal NEC damage is often followed by multisys- cells without involvement of progenitor cells. When the turn-
tem organ failure, including liver failure.4–6 The prognosis of over capacity of mature hepatocytes is overwhelmed by severe
patients with NEC is related to the development of multisystem liver injury, hepatic progenitor cells (HPCs) are recruited to
organ failure as it is often the cause of death.4,5 It has been regenerate and repair the liver.8–10 Katoonizadeh et al, reported
shown that liver injury is present in experimental NEC.6,7 that the number of HPCs was inversely correlated to the
Halpern et al, reported that the gut–liver inflammatory inter- number of Ki67þ hepatocytes in human acute liver failure.18
action is involved in the mechanism of NEC-associated liver We recently reported that the expression of proliferation
injury.6 Inflammatory cytokines produced in the intestine markers was significantly lower in experimental NEC liver
travel to the liver through the portal vein. At the same time, compared to control, suggesting that the proliferative capacity

Fig. 4 Expression of mRNA in liver organoids. (A) mRNA expression of liver progenitor cell markers LGR5 and SOX9. Expression of liver progenitor
cell markers were significantly lower in NEC compared to control (LGR5: p < 0.05; SOX9: p < 0.05). (B) mRNA expression of liver proliferation
markers Ki67 and PCNA. Expression of proliferation markers were significantly lower in NEC compared to control (Ki67: p < 0.05; PCNA:
p < 0.05). NEC, necrotizing enterocolitis; mRNA, messenger ribonucleic acid; PCNA, proliferating cell nuclear antigen; LGR5, leucine rich repeat
containing G protein-coupled receptor 5.
is impaired in the liver during the acute stage of the disease.12 In injury such as NEC, if exposed to various stress factors.20
such situation, HPCs are expected to play a role in liver These studies implicate that organoids can be useful as
regeneration and LGR5 is considered the marker for liver disease model of neonatal diseases.
progenitor cells.8 Khan et al., reported that the expression of
LGR5 positive cells was increased in pediatric chronic liver
Conclusion
disease, suggesting that LGR5 positive cells play a role in liver
regeneration.10 We reported that the expression of LGR5 in the We successfully derived organoids from the liver of neonatal
liver is impaired after NEC induction, suggesting that during mice with NEC. Compared to “normal liver organoids,” the
NEC, the liver regenerative capacity is impaired.12 “NEC liver organoids” grow slower and have less regenerative
In vivo animal model of NEC are used to test new capacity. This is consistent with what observed in the whole
therapeutic agents aimed to reduce the intestinal injury.19 liver,12 indicating that “NEC liver organoids” recapitulate the
LGR5 positive cells are known as progenitor cells of both changes of the entire organ of origin. By using this organoid
intestinal epithelium and hepatic hepatocytes and ductal platform, we can obtain a deeper understanding of the
cells. Therapeutic agents which have beneficial effects on mechanism of liver dysfunction in NEC and develop novel
intestine regeneration may also have positive effects in the therapeutic strategies.
liver, resulting in reduced NEC severity and lower liver
damage. Conflict of Interest
Organoids have similar features of their tissue of origin.13 None declared.
Organoids from various organs, such as stomach, intestine,
pancreas, liver, and prostate have been recently used as
disease model in cancer.13–15 In pediatrics, it has been References
1 Neu J, Walker WA. Necrotizing enterocolitis. N Engl J Med 2011;
reported that liver organoids from children with α1-anti-
364(03):255–264
trypsin deficiency and Alagille’s syndrome mirror the in vivo 2 Hall NJ, Eaton S, Pierro A. Royal Australasia of Surgeons Guest
disease.13 We reported that intestinal organoids derived Lecture. Necrotizing enterocolitis: prevention, treatment, and
from neonatal mice present similar features to an intestinal outcome. J Pediatr Surg 2013;48(12):2359–2367

3 Hull MA, Fisher JG, Gutierrez IM, et al. Mortality and management 12 Miyake H, Li B, Lee C, et al. Liver damage, proliferation, and
of surgical necrotizing enterocolitis in very low birth weight progenitor cell markers in experimental necrotizing enterocolitis.
neonates: a prospective cohort study. J Am Coll Surg 2014;218 J Pediatr Surg 2018;53(05):909–913
(06):1148–1155 13 Huch M, Gehart H, van Boxtel R, et al. Long-term culture of
4 Sonntag J, Wagner MH, Waldschmidt J, Wit J, Obladen M. Multi- genome-stable bipotent stem cells from adult human liver. Cell
system organ failure and capillary leak syndrome in severe 2015;160(1-2):299–312
necrotizing enterocolitis of very low birth weight infants. J Pediatr 14 Broutier L, Andersson-Rolf A, Hindley CJ, et al. Culture and
Surg 1998;33(03):481–484 establishment of self-renewing human and mouse adult liver
5 Morecroft JA, Spitz L, Hamilton PA, Holmes SJ. Necrotizing entero- and pancreas 3D organoids and their genetic manipulation. Nat
colitis–multisystem organ failure of the newborn? Acta Paediatr Protoc 2016;11(09):1724–1743
Suppl 1994;396:21–23 15 Broutier L, Mastrogiovanni G, Verstegen MMA, et al. Human prima-
6 Halpern MD, Holubec H, Dominguez JA, et al. Hepatic inflamma- ry liver cancer-derived organoid cultures for disease modeling and
tory mediators contribute to intestinal damage in necrotizing drug screening. Nat Med 2017;23(12):1424–1435
enterocolitis. Am J Physiol Gastrointest Liver Physiol 2003;284 16 Barlow B, Santulli TV, Heird WC, Pitt J, Blanc WA, Schullinger JN. An
(04):G695–G702 experimental study of acute neonatal enterocolitis–the importance
7 Halpern MD, Holubec H, Clark JA, et al. Epidermal growth factor of breast milk. J Pediatr Surg 1974;9(05):587–595
reduces hepatic sequelae in experimental necrotizing enterocoli- 17 Zani A, Cordischi L, Cananzi M, et al. Assessment of a neonatal rat
tis. Biol Neonate 2006;89(04):227–235 model of necrotizing enterocolitis. Eur J Pediatr Surg 2008;18(06):
8 Huch M, Boj SF, Clevers H. Lgr5(þ) liver stem cells, hepatic organoids 423–426
and regenerative medicine. Regen Med 2013;8(04):385–387 18 Katoonizadeh A, Nevens F, Verslype C, Pirenne J, Roskams T. Liver
9 Forbes SJ, Newsome PN. Liver regeneration - mechanisms and regeneration in acute severe liver impairment: a clinicopatholog-
models to clinical application. Nat Rev Gastroenterol Hepatol ical correlation study. Liver Int 2006;26(10):1225–1233
2016;13(08):473–485 19 Lu P, Sodhi CP, Jia H, et al. Animal models of gastrointestinal and liver
10 Khan Z, Orr A, Michalopoulos GK, Ranganathan S. Immunohisto- diseases. Animal models of necrotizing enterocolitis: pathophysi-
chemical analysis of the stem cell marker LGR5 in pediatric liver ology, translational relevance, and challenges. Am J Physiol Gastro-
disease. Pediatr Dev Pathol 2017;20(01):16–27 intest Liver Physiol 2014;306(11):G917–G928
11 Huch M, Dorrell C, Boj SF, et al. In vitro expansion of single Lgr5þ 20 Li B, Lee C, Cadete M, Miyake H, Lee D, Pierro A. Neonatal intestinal
liver stem cells induced by Wnt-driven regeneration. Nature organoids as an ex vivo approach to study early intestinal
2013;494(7436):247–250 epithelial disorders. Pediatr Surg Int 2019;35(01):3–7

Original Article
Gravitational Autoreposition for Staged Closure

of Omphaloceles
Marie Uecker1 Claus Petersen1 Carmen Dingemann1 Caroline Fortmann1 Benno M. Ure1
Jens Dingemann1
1 Center of Pediatric Surgery, Hannover Medical School and Bult Address for correspondence Marie Uecker, MD, Center of Pediatric
Children’s Hospital, Hannover, Germany Surgery, Hannover Medical School and Bult Children’s Hospital,
Carl-Neuberg-Str. 1, 30625 Hannover, Germany
Eur J Pediatr Surg (e-mail: [email protected]).
Abstract Aim Management strategies for large omphaloceles remain controversial. In this
study, we discuss the use of GRAVITAS (gravitational autoreposition sutures), the
method used at our institution when successful primary closure is deemed question-
able. Patient’s primary clinical course and long-term outcomes were analyzed.
Materials and Methods This is a single-center retrospective analysis of all consecutive
patients with omphaloceles treated between 1997 and 2018. Decision for GRAVITAS
was made when the defect was estimated too large for primary closure. Traction
sutures were placed in the fascia surrounding the defect and then suspended from the
top of the incubator to allow gravitational autoreposition of the herniated organs.
Ventilation and muscle relaxation were maintained until secondary closure, which was
performed after the obtruding viscera had been reduced by repeated adjustment of the
suture’s tension. Data are presented as mean standard deviation.
Results Out of 49 patients with omphaloceles, 12 were treated with GRAVITAS, 33
underwent primary closure, and 4 were treated using Schuster’s technique. Mean time
to secondary closure after GRAVITAS was 7 10 days. In nine of the patients who had
isolated omphalocele, secondary closure was achieved after 4 2 days. Ventilation
time was 5 2 days, and time to full feeds was 18 16 days. In three patients (one
with Fallot’s tetralogy, one with Cantrell’s pentalogy, and one with lung hypoplasia),
abdominal closure was achieved after 17 15 days. Due to cardiorespiratory comor-
bidity, ventilation time was >30 days. Five patients received initial closure of the skin
and secondary fascial closure after 18 15 months. One patient with prior fascial
closure underwent later repair of an abdominal wall hernia. During follow-up (30 35
Keywords months), one patient with gastrointestinal obstruction due to adhesions required
► omphalocele laparotomy, and one patient with gastroesophageal reflux disease underwent
► staged closure fundoplication.
► congenital abdominal Conclusion GRAVITAS is a feasible method for staged closure of large omphaloceles
wall defect when successful primary closure is deemed questionable.
Caroline Fortmann's ORCID is https://orcid.org/0000-0002-

2289-5330.

June 15, 2019
GRAVITAS for Staged Closure of Omphaloceles Uecker et al.
Introduction the umbilical cord or the amniotic sac,11,12 as well as

allogenic mesh inlays and skin flap techniques for covering
Omphaloceles are congenital midline abdominal wall defects the defects.13,14 We present a method of staged closure for
with herniation of abdominal organs into a membrane- omphaloceles used at our institution when primary closure
covered sac that fail to return to the abdominal cavity during is deemed questionable and retrospectively analyzed the
embryonic development. The goal of surgical reconstruction clinical course as well as the outcome of these patients
is to place the abnormally located viscera back in the including long-term follow-up.
abdominal cavity and recreate a functional abdominal wall
without organ injury or risking hemodynamic and respira-
tory compromise.
Smaller defects may be safely repaired with primary closure A retrospective chart review of patients with omphaloceles
soon after birth, whereas larger defects require more complex treated at the Hannover Medical School between 1997 and 2018
care due to the disproportion of herniated viscera and intra- was performed. Data collection included patient’s character-
abdominal space, which can lead to abdominal compartment istics, comorbidities, and method of treatment. For patients
syndrome when closed too quickly.1 Cases may be complicated with staged closure method of treatment, size of the defect and
by associated congenital malformations such as congenital herniated organs were recorded, as well as time to secondary
heart disease and lung hypoplasia.2 These comorbidities, if closure, time on ventilation, time to full feeds, and length of
present, require comprehensive multidisciplinary care and initial hospital stay. Time to full feeds was considered the
often have a severe impact on the outcome and prognosis of number of days until administration of parenteral nutrition
the affected patients.2,3 was stopped. Patients treated with gravitational autoreposition
Optimal management of large omphaloceles not suitable sutures (GRAVITAS) were scheduled for a clinical follow-up in
for primary closure remains controversial. The large variety the outpatient clinic within 6 to 12 weeks after discharge and
of surgical techniques described in the literature reflects the then seen at regular intervals depending on the clinical course.
challenge that these patients continue to represent for Patients with comorbidities were additionally followed up by
pediatric surgeons.4 When success of primary closure seems the pediatric department. Follow-up data collection was fo-
unlikely, the treating surgeons may choose between staged cused on additional abdominal surgical procedures and gastro-
and delayed closure, the latter involving the application of intestinal complications such as hernia repairs, intestinal
topical substances to promote membrane epithelization of obstruction, and feeding issues. The study was approved by
the sac and fascial closure later on5,6 (►Table 1). Since the the Institutional Review Board (#100519/DA).
placement of a silastic silo as a method of staged closure has
first been described by Schuster in 1967,7 numerous techni- Surgical Technique
ques for staged closure in the newborn period have been The standard treatment for patients at our institution born
published. These include the use of tissue expanders,8 com- with omphaloceles estimated too large for primary closure is
pression through wrapping or taping,9,10 and suspension of the placement of traction sutures to allow for gradual
Table 1 Advantages and disadvantages of the different surgical options for patients with omphaloceles1,4,6,7
Method of Technique Indication Advantage Disadvantages/

closure possible complications
Primary Immediate closure of the Small defects Only one operation Abdominal compartment
closure defect after birth needed syndrome
Delayed Application of topical Large defects Noninvasive Risk of sac rupture or
closure substances Patients not suitable for Allows time for growth infection
surgery/anesthesia in the before surgical Side effects of topical
neonatal period intervention agents used
Schuster’s Suture placement of a Large defects estimated Protection of herniated Multiple surgical
method silastic silo around the unsuitable for primary organs until secondary procedures
defect, removal of the Silo closure closure Risk of infection of the
before secondary closure Herniation of majority of Permanent secondary sac/fascia
the liver closure in the neonatal Disruption of the suture
Comorbidities with higher period line
risk for hemodynamic or
respiratory compromise
GRAVITAS Circumferential place- Large defects estimated Leaves the fascia, skin, Longer time on ventilation
ment of tension sutures in unsuitable for primary and sac intact and under general
the fascia, suspension with closure Low tension anesthesia
daily adjustment Herniation of the liver at secondary closure
Comorbidities with higher Allows for physiological
risk for hemodynamic or reconstruction of the
respiratory compromise abdominal wall

reduction of the herniated organs into the abdominal cavity tolerated. Patients are discharged as soon as they are on full
and stretching of the abdominal wall until tension- enteral feeds with sufficient weight gain and do not require
free secondary closure is possible. Decision for staged closure any further medical care for associated anomalies.
is made at the treating attending pediatric surgeon’s
discretion.
Results
After birth, the newborn is stabilized and checked for
associated congenital malformations. After induction of anes- Patient Characteristics
thesia and intubation, the umbilical cord is tied off with a Out of 49 patients treated for omphalocele at our institution
ligature, leaving the omphalocele sac intact. Thick, nonabsorb- over a period of 20 years, 12 were treated with GRAVITAS, 33
able traction sutures are placed circumferentially in the fascia received primary closure, and 4 were not considered suitable
surrounding the defect. The sac is then wrapped in sterile candidates for the GRAVITAS method and instead were treated
dressing, and sutures are clamped above the dressing and with conventional Schuster’s technique due to the following
suspended from the top of the incubator in a “parachutelike” circumstances: one patient had a rupture of the amniotic sac
fashion to allow for gravitational autoreposition of the herni- and tear of the liver capsule, one patient with Cantrell’s
ated organs and stretching of the abdominal wall (►Fig. 1). pentalogy had a very large defect that was treated with silo
Tension of the sutures is adjusted daily after rewrapping the placement and consecutive Gore-Tex patch closure, and two
sac under sterile conditions. Once the obtruding viscera have patients were considered unsuitable due to a severe underly-
been reduced into the abdominal cavity, the neonate is taken to ing syndromal disease with very limited prognosis.
the operating theater and secondary closure is performed. The In the 12 patients treated with GRAVITAS, mean gestational
tension sutures are removed, and the amniotic sac is resected. age at birth was 361/7 weeks 3 weeks and mean gestational
Possible adhesions with the liver are left intact to avoid liver weight was 2,690 490 g (n ¼ 7 males [58%]; n ¼ 5 females
injury and bleeding. After dissection of the abdominal wall [42%]). All patients were operated on within the first 24 hours
layers, closure of fascia and skin is performed with absorbable of life. Nine patients had omphalocele with no or minor
interrupted sutures. comorbidities not primarily affecting the outcome, and three
Patients are kept on ventilation and muscle relaxation for patients had severe comorbidities (one with Cantrell’s pental-
the duration of the treatment and are extubated as soon as ogy, one with Fallot’s tetralogy, and one with lung hypoplasia
possible after secondary closure. Antibiotics are administered and pulmonary hypertension). 90% of the patients had herni-
until 24 hours after successful secondary closure to prevent ation of the liver (►Table 2).
infection. Enteral feeding is commenced after secondary
closure once the first stool has been passed and gastric fluid Perioperative Outcome
retention has reduced to a minimum. Feeding is increased as For patients without severe comorbidities, time to secondary
quickly as possible according to the patient’s tolerance, and closure and mean hospital stay were shorter than for patients
parenteral nutrition is ceased once full enteral feeds are with congenital malformations (►Table 3). In 10 cases, only
two operations were needed to achieve secondary closure, the
first being the placement of the tension sutures and the second
being removal of sutures and closure of the abdominal wall. In
one patient with severe lung hypoplasia and pulmonary
hypertension, attempted closure with GRAVITAS method
was not successful. The patient suffered a secondary sac
rupture and wound infection and required multiple revisional
surgeries. With the use of additional skin incisions and a
silastic silo, secondary closure with a Gore-Tex patch was
achieved 40 days after birth. Another patient with Fallot’s
tetralogy required secondary placement of a silastic sheet due
to a nonrepositionable liver. This patient was on ventilation for
70 days due to the underlying heart disease; the two patients
with lung hypoplasia and Cantrell’s pentalogy were discharged
with home ventilation.
Long-Term Follow-Up
All GRAVITAS patients were seen at least once in our follow-up
clinic after discharge. Mean follow-up was 30 35 months.
Five patients had initial skin closure with secondary fascial
closure after 18 15 months. One patient underwent repair
of a midline abdominal wall hernia 36 months after secondary
Fig. 1 GRAVITAS technique. (A) Placement of traction sutures after
birth while leaving the amniotic sac intact. (B) Suspended traction
closure. Six (50%) patients required bilateral inguinal hernia
sutures with sterile dressing of the sac. (C) Same patient 2 weeks repair within the first year of life. One patient who had
after secondary closure. undergone additional open congenital diaphragmatic hernia

Table 2 Overview of patient’s clinical course
Herniated Time to secondary Comorbidities Complications

organs closure
Liver, bowel 2 days Congenital diaphragmatic hernia Bilateral inguinal hernia, gastrointestinal
obstruction due to adhesions after 5 months
Liver, bowel 2 days, skin closure – Bilateral inguinal hernia
Liver, bowel 5 days, skin closure VSD Bilateral inguinal hernia
Liver, bowel 6 days, skin closure Cantrell’s pentalogy Bilateral inguinal hernia
Bowel 2 days PFO –
Liver, bowel 6 days – Bilateral inguinal hernia
Liver, bowel 6 days, skin closure – Fundoplication for gastroesophageal reflux disease
Liver, bowel 3 days – –
Liver, bowel, 7 days, skin closure Fallot’s tetralogy Three operations, liver not reducible, intermittent
spleen silastic sheet placement before secondary closure
Liver, bowel, 40 days Lung hypoplasia, tracheomalacia, Secondary ruptured amniotic sac, wound infection,
pancreas pulmonary hypertension multiple revisions: additional incisions,
placement of a silastic silo, closure eventually
with Gore-Tex patch
Liver, bowel 2 days – –
Liver, bowel 5 days – Bilateral inguinal hernia
Abbreviations: PFO, patent foramen ovale; VSD, ventricular septal defect.
Table 3 Outcome of patients treated with GRAVITAS
Patients with minor Patients with major All patients treated with
comorbidities/isolated comorbidities (n ¼ 3) GRAVITAS (n ¼ 12)
omphalocele (n ¼ 9)
Time to secondary closure 4 2 days 17 15 days 7 10 days
Time on ventilation 5 2 days n ¼ 1: 70 days, two patients 11 19 days (n ¼ 10)
discharged with home
ventilation
Time to full feeds 18 16 days 62 30 days 29 26 days
Length of stay 35 21 days 147 53 days 63 53 days
Abbreviation: GRAVITAS, gravitational autoreposition sutures.
(CDH) repair suffered from a bowel obstruction 5 months after omphalocele treated in Germany. The authors stated that
omphalocele and CDH repair and required laparotomy for comparison of different management strategies of ompha-
release of adhesive ileus. Another patient suffered from loceles reported in the literature was difficult due to hetero-
gastroesophageal reflux disease and was treated with a geneity of used methods and clinical data. In a recent study
fundoplication 1 year after omphalocele repair. Both patients by Dutch authors,16 relevant studies concerning the
have since been in a good clinical condition without residual treatment of giant omphaloceles were identified and ques-
symptoms and nutritional/feeding issues. Remaining patients tionnaires were sent to the authors to inquire whether the
showed no gastrointestinal complications during follow-up. once-advocated surgical techniques were still being used or
No patient treated with GRAVITAS died during or after had been modified or even abandoned. Publications were
treatment. categorized into primary closure, staged closure, and delayed
closure. Almost half (42%) of the authors admitted to having
changed or stopped using their techniques after publication,
Discussion
with changes not being made in favor of one particular
Consensus regarding optimal management of large ompha- proven better technique. The fact that 50% of authors who
loceles is lacking. None of the suggested techniques for had published about staged closure techniques had changed
staged closure has been proven superior to others at this their method goes to show that even individual medical
point. Dingemann et al15 recently analyzed health insurance centers struggle to find a satisfying method of treatment to
data and confirmed an excellent outcome in patients with be consistently used.

Various authors used a type of suspension to produce a Conclusion

synergistic gravitational effect for organ repositioning and
abdominal wall extension in staged closure. Choice of mate- GRAVITAS is a feasible method of staged closure for omphalo-
rials and means of application show a high variance between celes when successful primary closure is deemed questionable.
authors, for example, Mehrabi et al17 described the intricate The method can also be used in patients with complex comor-
“camel-litter” method, Pacilli et al18 use a constructed silo bidities. In complicated cases in which GRAVITAS is not
with Prolene Mesh, and Morabito et al10 share their tech- successful, the method can be combined with other techniques
nique with suspension of the umbilical cord and a circum- to achieve secondary closure. However, comparative studies are
ferential elastic body bandage. Brown and Wright11 and mandatory to obtain more evidence-based data to determine
Kogut and Fiore9 utilize tape for wrapping to achieve down- outcomes of the different strategies.
ward compression of the herniated organs.
Reported cohort sizes vary from 5 to 18 patients, and Conflict of Interest
mean time to secondary closure ranges from 4 to None declared.
26 days.9–11,17,18 Mean time to enteral feeds ranges from 6
to 14 days, mean time on ventilation ranged from 5 to 8 days,
and mean hospital stay ranges from a minimum of 11 to a References
maximum of 72 days. 1 Marven S, Owen A. Contemporary postnatal surgical manage-
We confirmed that our method allows for fast secondary ment strategies for congenital abdominal wall defects. Semin
closure in patients without major comorbidities. It is also Pediatr Surg 2008;17(04):222–235
2 Roux N, Jakubowicz D, Salomon L, et al. Early surgical management
viable for neonates with severe associated congenital malfor-
for giant omphalocele: Results and prognostic factors. J Pediatr Surg
mations, but these cases generally require a longer time
2018;53(10):1908–1913
until secondary closure. Mean time to full feeds in our cohort 3 Akinkuotu AC, Sheikh F, Olutoye OO, et al. Giant omphaloceles:
was slightly longer than in other publications, which might in surgical management and perinatal outcomes. J Surg Res 2015;
part be due to maintained ventilation and muscle relaxation 198(02):388–392
during our treatment as well as different feeding protocols. 4 Mortellaro VE, St Peter SD, Fike FB, Islam S. Review of the evidence
on the closure of abdominal wall defects. Pediatr Surg Int 2011;27
Mean time of hospital stay of around 35 days in our series is
(04):391–397
similar to other reports. Most authors report no complications 5 van Eijck FC, de Blaauw I, Bleichrodt RP, et al. Closure of giant
during follow-up. The need for hernia repairs or fascial closure omphaloceles by the abdominal wall component separation
at a later point is hardly being addressed, except by Kogut and technique in infants. J Pediatr Surg 2008;43(01):246–250
Fiore,9 making this outcome difficult to compare. 6 Bauman B, Stephens D, Gershone H, et al. Management of giant
The longer clinical course in patients with associated mal- omphaloceles: a systematic review of methods of staged surgical
vs. nonoperative delayed closure. J Pediatr Surg 2016;51(10):
formations is a known fact, as such malformations represent
1725–1730
the most important prognostic factor in patients with giant 7 Schuster SR. A new method for the staged repair of large ompha-
omphaloceles.2,3 In these patients, several surgical steps may loceles. Surg Gynecol Obstet 1967;125(04):837–850
be required to achieve fascial closure,8,19,20 and very large 8 Adetayo OA, Aka AA, Ray AO. The use of intra-abdominal tissue
defects, as well as infection and suture line disruption repre- expansion for the management of giant omphaloceles: review of
literature and a case report. Ann Plast Surg 2012;69(01):104–108
sent the most frequent aggravating factors.14,21
9 Kogut KA, Fiore NF. Nonoperative management of giant ompha-
Aside from surgical sutures and sterile dressings, GRAVITAS
locele leading to early fascial closure. J Pediatr Surg 2018;53(12):
bears the advantage of not requiring additional or costly 2404–2408
materials. Except for one wound infection, neither suture line 10 Morabito A, Owen A, Bianchi A. Traction-compression-closure for
disruption nor respiratory or hemodynamic compromise was exomphalos major. J Pediatr Surg 2006;41(11):1850–1853
observed. In two patients, GRAVITAS was changed to Schuster’s 11 Brown MF, Wright L. Delayed external compression reduction of
an omphalocele (DECRO): an alternative method of treatment for
after GRAVITAS was not successful in achieving secondary
moderate and large omphaloceles. J Pediatr Surg 1998;33(07):
closure. Since GRAVITAS leaves the amniotic sac, fascia, and 1113–1115
skin intact, it is comparably noninvasive and allows for the 12 Hendrickson RJ, Partrick DA, Janik JS. Management of giant ompha-
addition of or transition to other techniques to achieve abdom- locele in a premature low-birth-weight neonate utilizing a bedside
inal wall closure (silo or patch placement, skin flap closure, etc.). sequential clamping technique without prosthesis. J Pediatr Surg
Limitations of our study are the retrospective nature, the 2003;38(10):E14–E16
13 Baird R, Gholoum S, Laberge J-M, Puligandla P. Management of a
small number of patients, and the lack of a control group, all
giant omphalocele with an external skin closure system. J Pediatr
of which do not provide enough evidence to make assertions. Surg 2010;45(07):E17–E20
The low incidence of large omphaloceles is partially respon- 14 Bawazir OA, Wong A, Sigalet DL. Absorbable mesh and skin flaps or
sible for the challenge of determining optimal treatment grafts in the management of ruptured giant omphalocele. J Pediatr
since comprehensive research and outcome analysis are Surg 2003;38(05):725–728
currently hard to realize with most publication’s patient 15 Dingemann C, Dietrich J, Zeidler J, et al. Surgical management of
congenital abdominal wall defects in germany: a population-
numbers being in the single digits. In addition, the constant
based study and comparison with literature reports. Eur J Pediatr
modification of techniques reported by van Eijck et al16 Surg 2017;27(06):516–525
further impedes the efforts to establish and compare differ- 16 van Eijck FC, Aronson DA, Hoogeveen YL, Wijnen RMH. Past and
ent methods, as also shown by Dingemann et al.15 current surgical treatment of giant omphalocele: outcome of a

questionnaire sent to authors. J Pediatr Surg 2011;46(03): 19 Kruit AS, Al-Ani SA, Jester I, Jester A. Multilayered flap technique:
482–488 a method for delayed closure of giant omphalocele. Ann Plast Surg
17 Mehrabi V, Mehrabi A, Kadivar M, Soleimani M, Fallahi A, Khalilzadeh 2016;76(06):680–683
N. Staged repair of giant recurrent omphalocele and gastroschesis 20 Parida L, Pal K, Al Buainain H, Elshafei H. Staged closure of
“camel-litter method”-a new technique. Acta Med Iran 2012;50(06): giant omphalocele using synthetic mesh. APSP J Case Rep 2014;
388–394 5(03):27
18 Pacilli M, Spitz L, Kiely EM, Curry J, Pierro A. Staged repair of giant 21 De Ugarte DA, Asch MJ, Hedrick MH, Atkinson JB. The use of tissue
omphalocele in the neonatal period. J Pediatr Surg 2005;40(05): expanders in the closure of a giant omphalocele. J Pediatr Surg
785–788 2004;39(04):613–615

Original Article
Lactoferrin Reduces Necrotizing Enterocolitis Severity

by Upregulating Intestinal Epithelial Proliferation
Jia Liu1,2 Haitao T. Zhu1,2 Bo Li2 Shaiya C. Robinson3 Carol Lee2 Joshua S. O’Connell2
Edoardo Bindi2 Shan Zheng1 Philip M. Sherman3 Agostino Pierro2
1 Department of Pediatric Surgery, Children’s Hospital of Fudan Address for correspondence Agostino Pierro, OBE, MD, FRCS, FAAP,
University, Shanghai, China Department of General and Thoracic Surgery, Hospital for Sick
2 Department of General and Thoracic Surgery, Hospital for Sick Children, 555 University Avenue, Toronto, ON M5G 1X8, Canada
Children, Toronto, Ontario, Canada (e-mail: [email protected]).
3 Department of Laboratory Medicine and Pathobiology, University of
Toronto, Ontario, Canada
Eur J Pediatr Surg
Abstract Introduction Necrotizing enterocolitis (NEC) is a devastating intestinal illness in

premature infants characterized by severe intestinal inflammation. Despite medical
interventions, NEC mortality remains alarmingly high, which necessitates improved
therapies. Lactoferrin is among the most abundant proteins in human milk and has
important immunomodulatory functions. While previous studies have indicated
protective effects of lactoferrin against neonatal sepsis and NEC, the underlying
mechanism remains unclear. We hypothesize that lactoferrin downregulates inflam-
mation and upregulates proliferation in intestinal epithelium during NEC injury.
Materials and Methods NEC was induced by hypoxia, gavage feeding of hyper-
osmolar formula and lipopolysaccharide between postnatal day P5 and P9 (n ¼ 8).
Breastfed mice were used as control (n ¼ 7). Lactoferrin (0.3 g/kg/day) was adminis-
tered once daily by gavage from P6 to P8 in both NEC (NEC þ Lac; n ¼ 9) and control
mice (Cont þ Lac; n ¼ 5). Distal ileum was harvested on P9 and analyzed for disease
severity, inflammation, and proliferation. Groups were compared using one-way
ANOVA and t-test appropriately; p < 0.05 was considered significant.
Results Compared to NEC group, lactoferrin-treated NEC mice had reduced disease
severity, reduced inflammation markers IL-6 and TNF-α expression and increased
Keywords intestinal stem cell marker Lgr5 þ expression. Lactoferrin-treated NEC mice exhibited
► lactoferrin increased nuclear β-catenin, indicating upregulated Wnt pathway, and increased Ki67
► necrotizing positivity, suggesting enhanced proliferation. Furthermore, lactoferrin administration
enterocolitis to control mice did not affect intestinal inflammation as well as Lgr5 þ stem cell
► inflammation expression and epithelial proliferation. This supports the safety of lactoferrin adminis-
► proliferation tration and indicates that the beneficial effects of lactoferrin are present when
► Wnt pathway intestinal injury such as NEC is present.
Bo Li's ORCID is https://orcid.org/0000-0002-8298-8344.

Agostino Pierro's ORCID is https://orcid.org/0000-0002-6742-
6570.

June 15, 2019
Lactoferrin Reduces Necrotizing Enterocolitis Severity Liu et al.
Conclusion Lactoferrin administration reduces the intestinal injury in experimental

NEC by downregulating inflammation and upregulating cell proliferation. This benefi-
cial effect of lactoferrin in stimulating cell proliferation is mediated by the Wnt
pathway. This experimental study provides insights on the mechanism of action of
lactoferrin in NEC and the role of lactoferrin in enteral feeding.
Introduction through P9, n ¼ 8; (3) NEC treated with lactoferrin (NEC þ

Lac): recombinant lactoferrin (0.3 g/kg/day, at a concentration
Necrotizing enterocolitis (NEC) is a severe disease that mostly of 6g/L; Sigma Aldrich, United States) administered once daily
develops in premature babies and has a high mortality estimat- by gavage from P6 to P8 in NEC mice, n ¼ 9. To evaluate
ed to be approximately 30%.1 While the precise etiology of NEC lactoferrin safety and investigate its effect in healthy mice,
is still unknown, prematurity, immune dysregulation, intesti- we further set up a group with recombinant lactoferrin
nal microbiota dysbiosis, and formula feeding are believed to be administered once daily by gavage from P6 to P8 in control
contributing factors. mice (Cont þ Lac), n ¼ 5. Distal ileum from all mice groups
Lactoferrin is an important protein consumed by breast-fed above were harvested on P9 and processed for further
new born infants, with the highest concentration in colostrum experiments.
(range 6–7 g/L) and a relatively lower concentration in mature
milk (range 1–5 g/L).2 As a multifunctional protein protective Quantitative Gene Expression
against infection and inflammation, lactoferrin has been used Total RNA was extracted from distal ileum using TRIzol
as a supplement for human infant formula.3 Lactoferrin is also reagent according to manufacturer’s instructions (Thermo
present in bovine milk at a concentration of 0.2 to 1.5 g/L with Fisher Scientific, Inc., IL, United States). cDNA was synthe-
over 70% amino acid homology compared to human lactofer- sized using qScript cDNA supermix (Quantabio, Beverly, MA,
rin.4,5 Therefore, both bovine lactoferrin and recombinant United States), and S1000 Thermal Cycler (Bio-Rad Labora-
human lactoferrin can be used in the human infant diet and tories, Inc., CA, United States). Quantitative real-time PCR
can supplement for the beneficial functions of human (qRT-PCR) was performed using the CF384 C1000 Thermal
lactoferrin. Cycler (Bio-Rad Laboratories, Inc., CA, United States) in SYBR
Previous studies have demonstrated a protective effect of Green medium system, as described previously.12 Interleukin
lactoferrin against infection and other serious diseases such 6 (IL-6) and tumor necrosis factor alpha (TNF-α) gene expres-
as NEC. Oral administration of bovine lactoferrin decreases sion were measured as inflammatory markers, and Leucine-
experimental colitis induced by dextran sulfate sodium.6 rich repeat-containing G-protein coupled receptor 5 (Lgr5)
Additionally, clinical studies of preterm infants have shown gene expression was measured as intestinal stem cell marker.
that bovine lactoferrin protects against late-onset sepsis,7 The expression of each gene was normalized to the house-
and both recombinant and bovine lactoferrin reduce the keeping gene Glyceraldehyde-3-Phosphate Dehydrogenase
incidence of Bell’s stage II and III NEC.8–10 Though several (GAPDH). Primer sequences were verified in nucleotide
studies investigated multiple aspects of lactoferrin function, BLAST database (►Table 1). Results were analyzed and
the underlying mechanism of its protective effect during NEC normalized to the control group. All assays were performed
progression remains unclear, and its protective effects on the in technical and experimental triplicate.
early stages of NEC is still controversial.2,11 Therefore, we
sought to elucidate the mechanisms driving lactoferrin-
mediated protection during NEC progression, especially Table 1 Primer genes used in running quantitative reverse
focused on its regulation of inflammation and cell prolifera- transcription polychain reaction (qRT-PCR) along with their
tion. Furthermore, we administered lactoferrin on healthy forward and reverse sequences
pup mice to verify its safety and test for adverse effects.
Primer Forward Reverse
GAPDH 5′-TGAAGCAGGCATC 5′-CGAAGGTGGAAG
Materials and Methods TGAGGG-3′ AGTGGGAG-3′
Experimental Groups and Sample Collection IL-6 5′-CCAATTTCCAATG 5′-ACCACAGTGAGG
All animal experiments were approved by the Animal Care CTCTCCT-3′ AATGTCCA-3′
Committee at The Hospital for Sick Children (no. 44.032), and TNF-α 5′-TTCCGAATTCACTG 5′-TGCACCTCAGGG
all methods were performed according to its guidelines and GAGCCTCGAA-3′ AAGAATCTGGAA-3′
regulations. C57BL/6 mice were purchased from The Jackson Lgr5 5′-CGAGCCTTACAGA 5′-TTGCCGTCGTCTT
Laboratory and kept in laminar-flow cages. We first set up GCCTGATACC-3′ TATTCCATTGG-3′
three experimental groups: (1) Control: breastfed mice, n ¼ 7; Abbreviations: GAPDH, glyceraldehyde phosphate dehydrogenase;
(2) NEC: induced by hypoxia, gavage feeding of hyperosmolar TNF-α, tumor necrosis factor-alpha; Lgr5, leucine-rich repeat-containing
formula, and lipopolysaccharide (LPS) from postnatal day P5 G-protein coupled receptor 5.

Hematoxylin and Eosin Staining and Statistical Analysis

Immunohistochemistry Data are presented as mean standard deviation (SD). As
Distal ileum tissue samples were fixed in 4% paraformalde- appropriate, one-way ANOVA was used for comparisons
hyde, embedded in paraffin and sectioned at 5 μm thickness, among control, NEC, and lactoferrin-treated NEC groups; t-
hematoxylin and eosin (H and E) staining was performed by test was used for comparisons between control and lacto-
standard procedures to evaluate intestinal injury severity or ferrin-treated control groups. A p-value < 0.05 was consid-
prepared for further experiments. ered as statistically significant. All analyses were performed
Immunohistochemistry (IHC) was performed by a two- using GraphPad Prism 6 software.
step EnVision/HRP technique (Vector Laboratories, Burlin-
game, CA, United States) according to the manufacturer’s
Results
instructions. Nuclear factor kappa-light-chain-enhancer of
activated B-cells (NF-κB) was used as a marker of inflam- Intestinal Injury and Inflammation
mation, nuclear β-catenin was used as an indicator of the To evaluate the effects of lactoferrin on NEC outcomes, lacto-
Wnt pathway activation, and Ki67 was used as a marker of ferrin was administered to NEC mice and disease severity was
cell proliferation. The following antibodies were used for assessed. Compared to control (breastfed) mice, NEC mice
IHC: rabbit antihuman/mouse NF-κB p65 monoclonal anti- exhibited villus sloughing, which was attenuated by lactoferrin
body (Cell Signaling Technology, United States) diluted to treatment (►Fig. 1A). IHC analysis showed reduced nuclear
1:800; rabbit antihuman/mouse β-catenin monoclonal an- positivity for NF-κB in lactoferrin-treated NEC mice compared
tibody (Santa Cruz Biotechnology, United States) diluted to with NEC group (►Fig. 1B), suggesting that lactoferrin treat-
1:500; and rabbit antihuman/mouse Ki67 polyclonal anti- ment attenuates NEC-induced intestinal inflammation. Indeed,
body (Abcam, United Kingdom) diluted to 1:500. All dilu- mRNA expression of inflammation markers IL-6 and TNF-α
tions were made in 1% bovine serum albumin (BSA). Cell were significantly elevated in NEC mice, which showed increase
nuclei were counterstained using hematoxylin. Negative in IL-6 and TNF-α, relative to control mice (►Fig. 1C, D). Notably
controls were conducted using 1% BSA lacking primary however, lactoferrin-treated NEC mice exhibited a reduction in
antibody. As activated forms of NF-κB (phospho-NF-κB) IL-6 and TNF-α, compared to NEC mice not receiving lactoferrin
and β-catenin (nonphospho-β-catenin) are only detectable (IL-6, 0.7511 0.5526 vs. 2.410 1.061, p ¼ 0.0003; TNF-α,
in cell nuclei, cells with brown-stained nuclei were defined 0.8296 0.7217 vs. 2.965 1.100, p < 0.0001), and were
as positive. indistinguishable from control breastfed mice (►Fig. 1C, D).
H and E and IHC staining images were acquired using
Nikon TE-2000 digital microscope (Nikon Instruments Inc., Intestinal Epithelium Proliferation
NY, United States) with Hamamatsu C4742-80-12AG camera NEC mice exhibited reduced Ki67-positivity than their control
(Hamamatsu Photonics K.K, Hamamatsu, Japan). Quantifica- counterparts, while lactoferrin-treated NEC mice presented
tion of Ki67 and β-catenin expressions were analyzed by with restored Ki67 positivity (NEC þ Lac 1.706 0.2314 vs.
positive areas relative to control group, using ImageJ NEC 0.6290 0.1005, p ¼ 0.0208; ►Fig. 2A , B). Since pro-
software. liferation of stem and progenitor cells is driven by the Wnt
Fig. 1 Disease severity and inflammation levels. (A) H and E staining for intestinal injury severity. (B) Inflammation marker NF-κB, activated NF-
κB shown as coexpression of brown positive staining with purple nuclei stain expressed in all layers and various kinds of cells (black arrow). (C,D)
mRNA expression levels of inflammation markers IL-6 and TNF-α. NEC, necritizing enterocolitis; NF- κB, nuclear factor-κB; H and E, hematoxylin
and eosin; TNF-α, tumor necrosis factor-alpha; Lac, lactoferrin.

Fig. 2 Cell proliferation levels. (A) Cell proliferation marker Ki67, shown as coexpression of brown positive staining with purple nuclei stain of
crypt cells at the bottom of the intestine (black arrow). (B) Quantification of Ki67 expression in IHC by positive area relative to control group.
(C) Cell proliferation marker β-catenin in Wnt pathway, activated β-catenin shown as coexpression of brown positive staining with purple nuclei
stain of crypt cells at the bottom of the intestine (black arrowhead). (D) Quantification of nuclear β-catenin expression in IHC by positive area
relative to control group. (E) mRNA expression levels of intestinal stem cell marker Lgr5. IHC, immunohistochemistry; mRNA, messenger
ribonucleic acid; NEC, necritizing enterocolitis; Lac, lactoferrin; Lgr5, leucine rich repeat containing G protein-coupled receptor 5.
signaling pathway, we next assessed nuclear β-catenin administration during NEC restored Lgr5 expression similar
staining as an indication of Wnt-pathway activation. to control levels (1.298 0.4121 vs. 0.3951 0.1499,
Compared to control mice, NEC mice presented with p ¼ 0.0006; ►Fig. 2E).
decreased nuclear β-catenin, which was restored in the Lastly, we compared the mRNA expression levels of our
lactoferrin-treated NEC mice (NEC þ Lac 1.177 0.0811 vs. tested inflammation and cell proliferation markers between
NEC 0.4147 0.1865, ►Fig. 2C, D). In agreement with these control and lactoferrin-treated control groups to ensure
findings, the expression of Wnt target gene and intestinal lactoferrin safety. Lactoferrin administration to control
stem cell marker Lgr5 (mRNA expression) was decreased mice did not upregulate inflammation markers IL-6 or
during NEC relative to control mice. However, lactoferrin TNF-α (►Fig. 3A) or Lgr5 expression (►Fig. 3B).
Fig. 3 mRNA of inflammation markers IL-6 and TNF-α (A) and intestinal stem cell marker Lgr5 (B) in control and lactoferrin-treated control mice
(p > 0.05). Cont þ Lac, control þ lactoferrin; Lgr5, leucine rich repeat containing G protein-coupled receptor 5; TNF-α, tumor necrosis factor-
alpha; Lac, lactoferrin; mRNA, messenger ribonucleic acid.

Discussion whether lactoferrin treatment ameliorated NEC-mediated

damage by stimulating intestinal proliferation. In our study,
This study indicates that lactoferrin enteral administration is lactoferrin-treated NEC mice had significant higher gene
beneficial during NEC development as it reduces intestinal expression of Lgr5 and increased Ki67, indicating higher
inflammation and promotes epithelial proliferation via the intestinal epithelial cell proliferation levels. Moreover,
activation of intestinal stem cells. lactoferrin-treated NEC mice had increased β-catenin
Human breast milk is known to be protective against the expressed in the nuclei of crypt cells. As β-catenin plays a
development of NEC.13 However, human breast milk is not central role in the Wnt pathway and only nonphosphory-
always available, and it is not unusual that premature infants lated β-catenin can enter the cell nuclei, we hypothesized
are given formula feeding to support growth. Lactoferrin is that during NEC, lactoferrin can upregulate Wnt pathway by
contained in human breast milk, particularly in the early activating β-catenin and by stimulating cell proliferation. In
stages of lactation.3 A systematic review of various clinical vitro studies showed that human lactoferrin can increase
studies indicates that lactoferrin can reduce mortality and proliferation in rat crypt cells and porcine intestinal epithe-
relieve the severity of sepsis and NEC in premature lial cells.4,17 Similarly, in agreement with our findings, an in
infants.2,7,8 However, the mechanism of action of lactoferrin vivo study in healthy neonatal piglets demonstrates that
remains unclear and it is possible that its effects can explain dietary bovine lactoferrin is beneficial by increasing intesti-
some of the benefits of human milk feeding. In addition, nal epithelial cell proliferation and by improving or protect-
lactoferrin can be given as supplement of formula feeding ing intestinal crypt and villous structure.5
when human breast milk is not available. Lastly, we verified the safety of lactoferrin in healthy
Multiple functions of lactoferrin have been investigated in individuals. Lactoferrin administration in control mice did
previous studies,4,14 and different concentrations and dosages not result in increased inflammation or up-regulation of Lgr5
have been demonstrated to have opposite effects. Low dose expression. This supports the concept that the beneficial
(0.1–1g/L) lactoferrin can activate cell proliferation and down- effects of lactoferrin are related to the presence of intestinal
regulate inflammation, while a higher dose (10g/L) can be damage such as that observed during NEC.
detrimental by inhibiting cell proliferation and upregulating In conclusion, we show that lactoferrin administration
inflammation.4 In the present study, we used a dosage of reduces the severity of intestinal injury in NEC by down-
recombinant lactoferrin (0.3 g/kg/day) at a concentration of regulating inflammation and up-regulating cell proliferation
6 g/L, to approximate the concentration in human colostrum. without adverse effects. This beneficial effect of lactoferrin in
We found that lactoferrin administration improved the sever- stimulating cell proliferation is mediated, in part, by the
ity of intestinal injury of experimental NEC, downregulated canonical Wnt pathway. These findings support the impor-
inflammation markers (IL-6, TNF-α, nuclear NF-kB), and tance of lactoferrin administration in neonates. Lactoferrin
significantly upregulated the expression of intestinal stem could be added to enteral feeds (human breast milk or formula)
cell and proliferation markers (Lgr5, Ki67, and nuclear β- as a supplement aiming to prevent the development and
catenin). These findings support the use of lactoferrin as progression of severe intestinal diseases such as NEC.
formula supplement to prevent the development of NEC.
Previous in vitro studies have demonstrated that lactoferrin Conflict of Interest
has direct intestinal immunomodulatory and antiinflammatory None declared.
actions by modulating immune cells and cytokine expres-
sion.15,16 Both human and bovine lactoferrin can downregulate
the LPS-induced levels of TNF-α, IL-1, IL-8, and IL-6 in IEC-6 and References
Caco-2 cells.14 Bovine lactoferrin of 0.1–1g/L administrated to 1 He YM, Li X, Perego M, et al. Transitory presence of myeloid-
NEC piglet can stabilize hypoxia-induced factor-1a (HIF-1α) and derived suppressor cells in neonates is critical for control of
inflammation. [J] Nat Med 2018;24(02):224–231
inhibit inflammation pathways including NF-κB, TNF receptor-
2 He Y, Cao L, Yu J. Prophylactic lactoferrin for preventing late-onset
associated factor 6 and p38 pathways.4 Lactoferrin administra- sepsis and necrotizing enterocolitis in preterm infants: a PRISMA-
tion at a high concentration (10 g/L) upregulated NF-κB compliant systematic review and meta-analysis. [J] Medicine
pathway and activated HIF-1α, resulting in colonic lesions (Baltimore) 2018;97(35):e11976
and reduced intestinal villus height/crypt depth ratio in the 3 Nguyen DN, Jiang P, Stensballe A, Bendixen E, Sangild PT, Chat-
immature small intestine of piglets.4 At 6 g/L, we detected terton DE. Bovine lactoferrin regulates cell survival, apoptosis and
inflammation in intestinal epithelial cells and preterm pig intes-
reduced nuclear NF-κB in lactoferrin-treated NEC mice com-
tine. [J] J Proteomics 2016;139:95–102
pared with NEC alone. This suggests that at this concentration, 4 Nguyen DN, Li Y, Sangild PT, Bering SB, Chatterton DE. Effects of
activation of NF-κB pathway is inhibited by lactoferrin admin- bovine lactoferrin on the immature porcine intestine. [J] Br J Nutr
istration, probably in relation to the down-regulation of the 2014;111(02):321–331
gene expression of inflammation markers IL-6 and TNF-α and 5 Reznikov EA, Comstock SS, Yi C, Contractor N, Donovan SM.
inhibition of downstream functional cytokines secretion. Dietary bovine lactoferrin increases intestinal cell proliferation
in neonatal piglets. [J] J Nutr 2014;144(09):1401–1408
Intestinal epithelial repair is driven, in part, by prolifera-
6 Togawa J, Nagase H, Tanaka K, et al. Oral administration of
tion of stem and progenitor cells at the base of the crypt. lactoferrin reduces colitis in rats via modulation of the immune
Another important property of lactoferrin is the stimulation system and correction of cytokine imbalance. [J] J Gastroenterol
of proliferation and differentiation.14 We sought to ascertain Hepatol 2002;17(12):1291–1298

7 Manzoni P, Rinaldi M, Cattani S, et al; Italian Task Force for the Study 12 Li B, Lee C, Martin Z, et al. Intestinal epithelial injury induced by
and Prevention of Neonatal Fungal Infections, Italian Society of maternal separation is protected by hydrogen sulfide. [J] J Pediatr
Neonatology. Bovine lactoferrin supplementation for prevention of Surg 2017;52(01):40–44
late-onset sepsis in very low-birth-weight neonates: a randomized 13 Lucas A, Cole TJ. Breast milk and neonatal necrotising enterocolitis.
trial. [J] JAMA 2009;302(13):1421–1428 [J] Lancet 1990;336(8730):1519–1523
8 Pammi M, Suresh G. Enteral lactoferrin supplementation for 14 Chatterton DE, Nguyen DN, Bering SB, Sangild PT. Anti-inflam-
prevention of sepsis and necrotizing enterocolitis in preterm matory mechanisms of bioactive milk proteins in the intestine
infants. [J] Cochrane Database Syst Rev 2017;6:CD007137 of newborns. [J] Int J Biochem Cell Biol 2013;45(08):
9 Manzoni P, Meyer M, Stolfi I, et al. Bovine lactoferrin supplemen- 1730–1747
tation for prevention of necrotizing enterocolitis in very-low- 15 Actor JK, Hwang SA, Kruzel ML. Lactoferrin as a natural immune
birth-weight neonates: a randomized clinical trial. [J] Early Hum modulator. [J] Curr Pharm Des 2009;15(17):1956–1973
Dev 2014;90(Suppl 1):S60–S65 16 Berlutti F, Schippa S, Morea C, et al. Lactoferrin downregulates
10 Meyer MP, Alexander T. Reduction in necrotizing enterocolitis and pro-inflammatory cytokines upexpressed in intestinal epithelial
improved outcomes in preterm infants following routine supple- cells infected with invasive or noninvasive escherichia coli strains.
mentation with lactobacillus GG in combination with bovine [J] Biochem Cell Biol 2006;84(03):351–357
lactoferrin. [J] J Neonatal Perinatal Med 2017;10(03):249–255 17 Nichols BL, McKee KS, Henry JF, Putman M. Human lactoferrin
11 Griffiths J, Jenkins P, Vargova M, et al. Enteral lactoferrin to stimulates thymidine incorporation into DNA of rat crypt cells. [J]
prevent infection for very preterm infants: the ELFIN RCT. [J] Pediatr Res 1987;21(06):563–567
Health Technol Assess 2018;22(74):1–60

Original Article
Clinical Factors Affecting Condition-Specific Quality-of-

Life Domains in Pediatric Patients after Repair of
Esophageal Atresia: The Swedish-German EA-QOL Study
Michaela Dellenmark-Blom1,2 Julia Quitmann3 Jens Dingemann4 Stefanie Witt3 Benno M. Ure4
Monika Bullinger3 Linus Jönsson2 Vladimir Gatzinsky2 Carmen Dingemann4
1 Department of Pediatrics, Institute of Clinical Sciences, University of Address for correspondence Michaela Dellenmark-Blom, RN, PCNS,
Gothenburg, Gothenburg, Sweden MSc, PhD, Department of Pediatric Surgery, The Queen Silvia
2 Department of Pediatric Surgery, The Queen Silvia Children’s Children’s Hospital, Rondvägen 10, 41685 Göteborg, Sweden
Hospital, Gothenburg, Sweden (e-mail: [email protected]).
3 Department of Medical Psychology, University Medical Center
Hamburg-Eppendorf, Hamburg, Germany
4 Department of Pediatric Surgery, Hannover Medical School,
Hannover, Niedersachsen, Germany
Eur J Pediatr Surg
Abstract Introduction We aimed to identify clinical factors affecting condition-specific health

related quality of life (HRQOL) domains in children born with esophageal atresia (EA).
This can facilitate preventive care to risk groups of HRQOL impairments.
Materials and Methods A total of 124 Swedish and German families of EA children
answered the validated EA-QOL questionnaires (response rate 68%), for evaluation of
three HRQOL domains in children 2 to 7 years old (53 parents) and four HRQOL domains
in children 8 to 17 years old (62 children/71 parents). Clinical data were collected
through medical records and a questionnaire. Statistics included between—group
analysis, univariable and stepwise multivariable regression analysis, p < 0.05.
Results Between 2 to 7 years, no primary anastomosis (p ¼ 0.022) and female gender
Keywords (p ¼ 0.026) predicted worse scores related to “physical health and treatment,” and
► esophageal atresia gastrostomy insertion related to “eating” (p ¼ 0.0001), and “social isolation and stress”
► quality of life (p ¼ 0.001). Between 8 to 17 years, no primary anastomosis (child report), prematurity,
► health-related quality esophageal dilatation (parent report) predicted poor HRQOL related to “eating”
of life (p < 0.05), associated anomalies to “body perception” (p ¼ 0.031, parent report), female
► pediatric surgery gender (p ¼ 0.018, child report) and severe EA (p ¼ 0.011 child report, p ¼ 0.004 parent
► children report) to “social relationships,” and severe EA predicted worse “health and well-being”
Introduction
quality of life (HRQOL). Following surgical correction, between
Esophageal atresia (EA) is a rare malformation characterized 56% and 76% of pediatric patients with EA suffer from dyspha-
by discontinuity of the esophagus, with or without a trache- gia or gastroesophageal reflux disease (GERD),3–5 and between
oesophageal fistula (TEF).1 As the survival in infants born with 56% and 69% from respiratory disorders like wheezing, cough,
EA has reached over 90%,2 increased attention is now being recurrent airway infections, and pneumonia.4–6 In 55% of
paid to the survivors’ long-term morbidity and health-related cases, EA is associated with concomitant anomalies, most
often related to the cardiovascular, urogenital, anorectal, and
Stefanie Witt's ORCID is https://orcid.org/0000-0002-6143-7594. vertebral-skeletal systems.7 Measurement of HRQOL includes

March 12, 2019 Stuttgart · New York 10.1055/s-0039-1693729.
June 15, 2019
Factors Affecting QOL in Pediatric Patients Dellenmark-Blom et al.
scores (p ¼ 0.004, parent report). An increased number of digestive symptoms (difficulty

swallowing food, heartburn, and vomiting), lowered all EA-QOL domain scores in both age
groups (p < 0.001). An increased number of respiratory problems (cough, wheezing,
airway infections. breathlessness, and chest tightness), lowered scores in two HRQOL
domains among children 2 to 7 years (p < 0.05).
Conclusion Impairments within condition-specific HRQOL domains in EA children are
found in congenital and surgical subgroups, and notably related to digestive symptoms
throughout childhood.
patients’ perceptions of the impact of disease and treatment on Hospital in Gothenburg, Sweden (n ¼ 86), the Centre of
their physical, social, and psychological functioning and well- Pediatric Surgery, Hannover Medical School and Bult Child-
being. While generic HRQOL assessments enable comparison ren’s Hospital, Hannover, Germany (n ¼ 102). For inclusion,
of outcomes between patient groups or with general norms, a participants needed to understand Swedish or German and
condition-specific approach captures content of importance to families should not have participated in the item generation
a particular disorder and clinical practice.8 It is important and pilot-testing of the EA-QOL questionnaires.15,16 Children
to identify EA patients who risk HRQOL restrictions, in order to of 8 years or older provided self-reports, and children
provide them timely care and treatment. younger than 8 years and those with cognitive impairments
Known risk factors for impaired generic overall HRQOL in EA were represented by parent-proxy report. Informed consent
children are prenatal diagnosis,9 EA Gross A,9 prematurity,7,9,10 was obtained from 182 families.
low birth weight,9 male gender,9 and severe EA.9 The digestive
risk factors are GERD,10 hospitalization due to esophageal Data Collection
illness, a gastrostomy insertion,11 and a long-term follow-up Families were contacted by a local researcher and received oral
for nutritional support.12 Respiratory risk factors are a barking and written study information. After the family gave study
cough, dyspnea on exertion,10 asthma, pneumonia, and hospi- approval, medical records were reviewed for clinical data and
talization due to respiratory illness.11 Associated anomalies, questionnaires were sent home to the family with prestamped
GERD and higher age13 affect general or physical health envelopes for their replies. They received a maximum of three
negatively, whereas patients with delayed esophageal repair reminders, to ensure good response rates. Clinical data collec-
have better levels of physical well-being than patients with tion included prenatal diagnosis of EA, gestational week, birth
esophageal replacement, major revisional surgery, or esoph- weight, child gender, associated anomalies, VACTERL, type of
ageal stenosis requiring >10 dilatations.14 Dysphagia and surgical correction, revisional surgery due to anastomotic
dyspnea are related to impaired emotional functioning, and leakage or recurrent TEF, gastrostomy insertion, esophageal
associated anomalies with impaired school-functioning.10 dilatation, and identified children with severe EA, using pre-
In recent years, condition-specific HRQOL instruments have defined criteria as described earlier,9,16,18 and in ►Table 1.
been developed and validated for EA patients aged 2 to 7 and 8 A parent-reported questionnaire developed by the
to 17 years.15–17 These instruments evaluate the age-specific authors was used to obtain information on sociodemo-
HRQOL domains as presented in ►Fig. 1. Clinical risk factors for graphic factors and the child’s digestive and respiratory
impairments within these HRQOL domains still need to be symptoms during the previous 4 weeks. Condition-specific
identified. This study aimed to identify clinical factors affecting HRQOL was measured by the EA-QOL questionnaires,17
condition-specific HRQOL domains in pediatric patients after which were developed in accordance with international
repair of EA. The main hypothesis of this study was that clinical standards for patient-reported outcomes.19–21 The domains
factors related to birth characteristics, surgery, and current EA- of the 17-item questionnaire for children 2 to 7 years (parent
related symptomology and indicate severity, would affect report) and of the 24-item questionnaire for children 8 to
these HRQOL domains negatively. 17 years (child- and parent report) are presented in ►Fig. 1.
These versions have all showed sound content, convergent,
concurrent, and known-groups validity, as well as satisfac-
tory retest reliability and internal consistency for all domains
The study was a part of a Swedish–German project focusing (Chronbach’s alpha coefficient 0.72–0.91). In ages 8 to
on HRQOL in children and adolescents born with EA and their 17 years, child- and parent report of the EA-QOL domains
families,9,15–17 and was approved by the Ethical Review demonstrate adequate level of agreement. All questions were
Boards of Gothenburg, Sweden (DNR 958-13) and Hannover, answered using a 4-week recall period, and a five-point
Germany (2936-2015). Likert’s scale from never (1) to always (5).17
Patients Statistical Analysis

Families of children born with EA Gross A-E aged 2 to 17 years Data were analyzed using SPSS 22.0 and SAS 9.software. The
(n ¼ 188) were recruited from the Queen Silvia Children’s distribution of continuous variables as mean and standard

Fig. 1 Illustration of the names and the main content of the health-related quality of life domains measured by the EA-QOL questionnaires for
children with repaired esophageal atresia, 2 to 7 and 8 to 17 years old.
Table 1 Children with severe esophageal atresia fulfilled at

and pediatric surgical perspectives (see clinical data collection),
least one of the following conditions 1 to 4
predictors resulting in p < 0.1 in a univariable analysis were
included in a linear stepwise multivariable regression analysis.
1 Delayed anastomosis or esophageal replacement
Furthermore, linear regression analysis was used to determine
2 Revisional surgery following anastomotic leakage the impact of the number of current digestive or respiratory
or recurrent fistula
symptoms on the EA-QOL domain scores.
3 Severe tracheomalacia, verified through
bronchoscopy as 75% collapse or excessive/severe
4 Another congenital anomaly resulting in disability Results
Study Population
A total of 124 families (68%) responded to the EA-QOL
deviation (SD) was given, while that of categorical variables was questionnaires (73/80 Swedish; 51/102 German). ►Table 2
n%. All statistical tests were conducted at the 0.05 significance presents the sample characteristics. There were no signifi-
level. Respondents and nonresponders were compared with cant differences between respondents vs. nonrespondents
regard to birth characteristics and pediatric surgical data as with regard to birth characteristics, gender, or pediatric
described earlier, using Fisher’s exact test. Responses to the surgical variables in either age group (p > 0.05).
EA-QOL questions were coded 1 to 5, where higher points
represented better HRQOL. The items were linearly transformed Differences in EA-QOL Domain Scores between Clinical
to a 0 to 100 domain scale, which required 70% item responses Subgroups
within each domain. For comparisons of the EA-QOL domain ►Supplementary Material (available in the online version)
scores between two independent subgroups, the Mann–Whit- details the EA-QOL domain scores in clinical subgroups of
ney U-test was used and effect sizes were used to estimate the children aged 2 to 7 years (parent report) and 8 to 17 years
magnitude and clinical meaningfulness between these groups (child and parent report), including p-values and effect sizes.
differences, applying Cohen’s d for a standardized interpreta- In children aged 2 to 7 years, prenatal diagnosis of EA and
tion; small (> 0.2), moderate (> 0.5), and large (> 0.8).22 In order gastrostomy insertion were associated with lower scores on
to determine predictors of EA-QOL domain scores from neonatal the domains “eating” and “social isolation and stress,”

Table 2 Presentation of the children born with esophageal atresia and their parents in the study sample
Children 2–7 years old Children 8–17 years old

(N ¼ 53) (N ¼ 71)
N (%) N (%)
Prenatal, congenital and surgical variables
Prenatal diagnosis of EA 4 (7.5) 7 (11.1)a
Male gender 35 (66.0) 40 (56.3)
Prematurely born (< 37 weeks) 24 (47.1) 25 (39.1)
Low birth weight (< 2,500 grams) 22 (43.1) 26 (41.9)
Associated anomalies 28 (52.8) 49 (69.0)
Cardiovascular anomalies 12 (22.6) 27 (38.0)
Anorectal anomalies 7 (13.2) 11 (15.5)
VACTERLb 9 (17.0) 13 (18.3)
Primary esophageal repair 47 (88.7) 60 (84.5)
Revisional surgery due to anastomotic 5 (9.4) 8 (11.3)
leakage or recurrent fistula
Gastrostomy insertion 14 (26.9) 21 (29.6)
1 esophageal dilatation(s) 15 (28.3) 32 (45.1)c
3 esophageal dilations 8 (15.1) 14(20.0)d
Severe EA 24 (45.3) 37 (52.1)
Digestive symptoms the past month
Difficulty swallowing food 22 (41.5) 20 (28.7)
Heartburn 15 (28.3) 22 (31.0)
Vomiting 16 (30.2) 11 (15.5)
Respiratory symptoms the past month
Wheezing 20 (37.7) 13 (18.3)
Cough 32 (60.3) 26 (36.6)
Airway infection 24 (45.3) 13 (18.3)
Breathlessness at rest/physical activity 16 (30.2) 22 (31.0)
Chest tightness 9 (17.0) 7 (9.9)
Parent report
Mother 48 (90.6) 64 (91.4)
Age (years), median (range) 37 (26–55) 45 (30–69)
University/college education 21 (39.6)c 34 (49.3)c
Abbreviation: EA, esophageal atresia.

a
8 Missing data.
b
VACTERL stands for vertebral defects, anal atresia, cardiac defects, tracheo-esophageal fistula, renal anomalies, and limb abnormalities. People
diagnosed with VACTERL association typically have at least three of these characteristic features.
c
2 Missing data.
d
2 Missing data.
whereas female children with EA had poorer scores in the (p ¼ 0.017), and also gastrostomy insertion (p ¼ 0.029), were
domain “physical health and treatment” than male children factors related to lower scores in the “body perception” domain
(p < 0.05, moderate to large effect sizes). (parent report). Prematurely born children with EA, children
In children aged 8 to 17 years, female gender (child report, who did not have primary anastomosis, or children who had
p ¼ 0.044), associated anomalies (child report, p ¼ 0.021), experienced at least one esophageal dilatation showed lower
specifically cardiovascular anomalies (child report, scores in the “eating” domain (child- and parent report,
p ¼ 0.039), and moreover major revisional surgery (parent p < 0.05). Moreover, prematurity, no primary anastomosis
report, p ¼ 0.037) were associated with lower scores in the (child– and parent report, p < 0.05), major revisional surgery
“social relationships” domain. Concomitant anomalies (child report, p ¼ 0.043), and gastrostomy insertion (child
(p ¼ 0.022), again particularly cardiovascular anomalies report, p ¼ 0.029; parent report, p ¼ 0.033) were all associated

with impaired “health and well-being.” Finally, and as detailed this population.15,16 The children’s HRQOL as measured in these
in ►Supplementary Material (available in the online version), domains is affected by congenital and surgical characteristics,
severe EA had a significant relationship to lower scores on all and across child ages prominently by digestive symptoms.
EA-QOL domains in child reports, and in all but “body percep-
tion” in parent reports, p < 0.05. Significant findings in this age Clinical Factors Affecting Condition-Specific HRQOL
group revealed several moderate effect sizes. Domains in 2 to 7 Year Old Children
In line with previous results showing impaired generic HRQOL
Independent Clinical Predictors of Impaired EA-QOL in EA children, mean child age 13.3 years10 or 7.7 years,11
Domain Scores respiratory symptoms negatively influenced condition-specific
►Table 3 lists the identified clinical predictors of HRQOL as HRQOL with regard to “physical health and treatment” and
measured by the EA-QOL domains, according to a stepwise “social isolation and stress” in EA children of 2 to 7 years. In this
multivariable regression analysis. study, the impact of airway symptoms on condition-specific
HRQOL was more apparent in children younger than 8 years
Influence of Digestive and Respiratory Symptoms on than older children. Possible explanations are that respiratory
the EA-QOL Domain Scores disorders in our study sample were more frequent in the
►Table 4 presents the impact of digestive and respiratory younger study cohort (►Table 2), in agreement with earlier
symptoms on scores in the EA-QOL domains according to a studies.23 Main risk groups of different condition-specific
linear regression analysis. As shown, an increased number of HRQOL impairments in the younger cohort were also patients
digestive symptoms lowered scores in all EA-QOL domains with no primary anastomosis and those who had received a
(child and parent report), p < 0.05. An increased number of gastrostomy insertion. These factors are both clinical indica-
respiratory problems decreased the EA-QOL domain scores in tions of EA severity. Gastrostomy insertion may be needed due
children 2 to 7 years old, affected only the “eating” domain in to long-gap EA, revisional surgery, cyanotic spells, severe GERD,
children 8 to 17 years old (p < 0.05), and explained generally or associated anomalies, and implies a severe clinical course.
less in the variation of scores than digestive symptoms. Therefore, it is likely that it is not the gastrostomy itself that
impacts HRQOL, but that it is referencing a severely diseased
subgroup of EA patients. At our centers, the main indications
Discussion
additional to long-gap EA, were severe nutrition difficulties,
This is the first study to describe clinical factors affecting severe GERD, or severe tracheomalacia, with risk of failure to
condition-specific HRQOL domains in an international sample thrive. Furthermore, our study results may be explained by the
of pediatric patients after repair of EA. These HRQOL domains age-specific questions composing the questionnaires,
have previously been shown to be particularly important for since these were developed from a child developmental
Table 3 The independent predictors of impaired condition-specific quality of life domains in children with repaired esophageal
atresia
Predictor β0 β1 R2 p-value
Children 2–7 years old (parent report)
Eating Gastrostomy insertion 81.2 25.3 0.25 0.0001
Physical health and treatment No primary esophageal repaira 70.7 22.4 0.16 0.022
Female gender 14.5 0.026
Social isolation and stress Gastrostomy insertion 84.9 21.9 0.21 0.001
8–17 years old (child report)
Eating No primary esophageal repaira 75.7 20.6 0.13 0.005
Social relationships Female gender 86.0 10.7 0.17 0.018
b
Severe esophageal atresia 11.5 0.011
8–17 years old (parent report)
Eating At least one esophageal dilatation 82.8 17.1 0.24 0.002
Gestational weeks at birth < 37 weeks 13.1 0.019
Social relationships Severe esophageal atresiab 84.0 12.6 0.12 0.004
Body perception Associated anomaly 88.3 11.1 0.07 0.031
b
Health and well-being Severe esophageal atresia 87.4 13.1 0.12 0.004
a
Delayed primary anastomosis or esophageal replacement.
b
Children with at least one of the following conditions: delayed anastomosis or esophageal replacement, major revisional surgery, severe
tracheomalacia, verified through bronchoscopy as 75% collapse or excessive/severe, another congenital disabling anomaly.

Table 4 The impact of digestive symptomsa and respiratory symptomsb on the EA-QOL domain scores in children with repaired
esophageal atresia aged 2 to 7 (parent report) and 8 to 17 years (child and parent report)
Digestive symptomsa Respiratory symptomsb

2–7 years old (parent report) β0 β1 R2 p-value β0 β1 R2 p-value
Eating 90.3 14.1 0.44 <0.001 82.2 4.2 0.09 0.050
Physical health & treatment 77.7 12.4 0.33 <0.001 78.8 7.7 0.30 <0.001
Social isolation & stress 91.1 9.8 0.27 <0.001 88.3 4.2 0.12 <0.05
8–17 years old (child report)
Eating 86.3 14.4 0.58 <0.001 77.4 4.7 0.07 <0.05
Social relationships 84.2 8.9 0.24 <0.001 76.4 0.8 0.002 0.7
Body perception 92.1 10.7 0.23 <0.001 85.8 3.5 0.03 0.2
Health and well-being 91.4 12.1 0.35 <0.001 85.0 4.5 0.06 0.07
8–17 years old (parent report)
Eating 86.9 13.8 0.43 <0.001 78.0 5.4 0.08 <0.05
Social relationships 86.9 8.9 0.27 <0.001 78.6 1.0 0.003 0.6
Body perception 89.9 9.4 0.25 <0.001 82.9 2.3 0.02 0.3
Health and well-being 93.4 12.9 0.51 <0.001 84.4 3.7 0.05 0.09
a
Digestive symptoms (nmax ¼ 3): food impaction/difficulty swallowing food, heartburn, vomiting.
b
Respiratory symptoms (nmax ¼ 5): cough, wheezing at physical activity/at rest, airway infections, breathlessness on physical exertion/at rest, chest
tightness.
perspective.15 They therefore, include questions regarding compared to previous reports.10,23,28 Although its aim is to
young EA children’s physical abilities to play sport and play, relief the patient’s symptom burden, it reflects group of
the respiratory burden, and problems finishing a whole meal patients with stricture-related and dysphagic problems, which
(►Fig. 1). Moreover, we observed that following repair of EA, can explain our findings.
girls in ages 2 to 7 years had poorer scores in the domain of Moreover, major revisional surgery in children 8 to
physical health and treatment, despite being clinically compa- 17 years old was associated with impaired HRQOL levels
rable with males in the same age group. Gender analysis in in the domains “eating”, “social relationships” and “health
HRQOL studies of this population are lacking24 and it is difficult and well-being.” The prevalence of major revisional surgery
to find a definite explanation of the results. Possibly, following in our study sample is in line with previous reports.29,30
repair of EA, girls in early childhood react differently to boys to Koivusalo et al29 observed that following major revisional
their physical function, or are treated differently than boys by surgery, 92% of EA patients eventually resumed full oral
parents and people near to them. intake of foods and those remaining dependent on gastro-
stomy feeding often had severe syndromes. In another
Clinical Factors Affecting Condition-Specific HRQOL study, having major revisional surgery was an independent
Domains in 8 to 17-Year Old Children predictor of the use of nutritional and educational support
In 8 to 17-year old children, several clinical factors affected the in school among EA children without genetic disorders.31
condition-specific HRQOL domains, though some of these Uniquely, the EA-QOL questionnaires gather multidimen-
factors may be interrelated. This could partly explain why, sional information on the children’s eating situation, and on
for example, prematurely born EA children and those with no their experiences of loneliness, teasing, and impaired com-
primary anastomosis both had lower scores of “eating” and munication due to EA (►Fig. 1). The study findings may
“health and well-being”. Though, following the stepwise reflect that the EA-QOL questionnaires sensitively captured
regression analysis, the variable “no primary anastomosis” the HRQOL impact in a severely diseased subgroup of EA
explained the level of “eating-QOL” in EA children in this age children. However, revisional surgery was not a predictor
group. Since it signifies a severe malformation with prevalent according to the stepwise regression analysis. Instead, a
and complex morbidity, its relationship to impaired “eating- main risk factor for impaired scores in several condition-
QOL” in children with EA could be expected.23,25 Still, few specific HRQOL domains in this age group was “severe EA”.
studies of generic HRQOL9 or of feeding difficulties23,26,27 have This category was established through international pedi-
demonstrated that long-gap EA or time to primary anastomo- atric surgical expertise, HRQOL methodology expertise and
sis play a role in the development of future problems. In previous research.5,14,29,32,33 It defined EA severity by
children of 8 to 17 years, at least one esophageal dilatation taking into account congenital and early complications,
was a main risk factor for lower “eating-QOL.” The prevalence while holistically addressing esophageal, respiratory, and
of esophageal dilatations is slightly lower in our study sample other disabling morbidities which hypothetically influence

HRQOL. Hence, the definition of severe EA can help to assess and that psychosocial HRQOL decreases with increasing age
the risk for impaired HRQOL in this age group.9 in EA children.12 This study provides details of which HRQOL
In this study, girls with repaired EA aged 8 to 17 years domains these restrictions might relate to, and defines
were negatively affected in relation to their social relation- patients at risk of impaired condition-specific HRQOL, with
ships compared to boys of the same age, with comparable mostly moderate to large clinical meaningfulness. Based on
clinical characteristics. As mentioned previously, gender these findings, supportive and targeted interventions should
analysis in EA studies are sparse.24 In a recent study,9 we be provided to the risk groups. Areas for future research
used the generic PedsQL 4.0 HRQOL questionnaire and found include the impact of different surgical techniques, child
that after EA repair, Swedish and German boys aged 2 to gender, coping strategies, socioeconomic standards, culture,
18 years had lower overall HRQOL. The differences in gender and psychoeducational interventions on HRQOL. Continuing
outcomes between generic and condition-specific HRQOL clinical analysis using the EA-QOL questionnaires in addi-
findings may be attributable to the measurement level. tional countries would also be valuable.
Alternatively, people’s perception of HRQOL may depend
on environmental and sociocultural norms.34,35 If the norms
Conclusion
of girls and boys born with EA differ, it could explain the
findings, though, future studies on this topic are needed. The condition-specific HRQOL domains in pediatric patients
Associated anomalies were related to impaired “body per- after EA repair are negatively and noticeably affected by
ception”. Depending on their particular defect, some children digestive symptoms throughout childhood, whereas respira-
undergo repeated surgical interventions. Cardiovascular tory symptoms mainly affect children between 2 and 7 years.
anomalies are also related to impaired “body perception” Female children born with EA risk negatively affected physi-
and the Swedish study center is one of two national centers cal health and treatment when young, and impaired social
for pediatric cardiac surgery, contributing to a slightly higher relationships in school age and early adolescence. The risk
rate of such anomalies than usually reported.36 VACTERL was factors are both congenital and surgical in nature. Between 2
not a risk factor of condition-specific HRQOL, which may to 7 years, these main factors are; no primary anastomosis
indicate that it is not the number of malformations, but the (“physical health and treatment”) and gastrostomy insertion
severity of a malformation that matters for HRQOL. Perhaps (“eating”; “social isolation and stress”). Between 8 to 17 years
because the EA-QOL questionnaires focus on issues related to the main risk factors include; prematurity, no primary
upper gastrointestinal and respiratory problems, anorectal anastomosis, esophageal dilatation (“eating”), associated
malformations did not influence condition-specific HRQOL. anomalies (“body perception”) and severe EA (“social rela-
tionships”; “health and well-being”). The risk groups should
Study Strengths and Limitations be provided targeted health care interventions, which pro-
This study excluded families of EA children who participated mote them to achieve good HRQOL.
in the item generation and pilot-testing of the EA-QOL
questionnaires.15,16 However, the study sample reflects the Conflict of Interest
answers of those who participated in the field test of the EA- M.D.B. reports grants from The Jerring Foundation and the
QOL questionnaires.17 Although this may be a limitation, the Swedish state under the agreement between the Swedish
goal was to avoid burdening the families by collecting new data. government and the country councils, the ALF-agree-
The international response rate was acceptable, but differed ment, DNR 717911, during the conduct of the study. All
between Sweden and Germany even with standardized meth- other authors reported no conflict of interest.
ods. A possible explanation is that the Swedish centralized care
of EA patients, different from the decentralized German sys-
tem, might have strengthened the family-caregiver relation- References
1 Pedersen RN, Calzolari E, Husby S, Garne E; EUROCAT Working
ship and families’ participation in the study. However, our
group. Oesophageal atresia: prevalence, prenatal diagnosis and
centers provide structured follow-up programs and the inter-
associated anomalies in 23 European regions. Arch Dis Child
national sample reinforces generalizability of the study results. 2012;97(03):227–232
Moreover, we selected a limited set of main clinical factors for 2 Zimmer J, Eaton S, Murchison LE, et al. State of play: eight decades of
the multivariable regression analysis, taking into consideration surgery for esophageal atresia. Eur J Pediatr Surg 2019;29(01):39–48
appropriate sample sizes for statistics. It should be noted that 3 Acher CW, Ostlie DJ, Leys CM, et al. Long-term outcomes of
patients with tracheoesophageal fistula/esophageal atresia: sur-
the predictors only partly explained levels of condition-specific
vey results from tracheoesophageal fistula/esophageal atresia
HRQOL, and do not necessarily reflect a causal relationship. It online communities. Eur J Pediatr Surg 2016; 26(06):476–480
strengthens the study that effect sizes provided information 4 Pedersen RN, Markøw S, Kruse-Andersen S, et al. Esophageal
about the clinical meaningfulness of the results. atresia: gastroesophageal functional follow-up in 5-15 year old
children. J Pediatr Surg 2013;48(12):2487–2495
Implication for Clinical Practice and Research 5 Olbers J, Gatzinsky V, Jonsson L, et al. Physiological studies at
7 years of age in children born with esophageal atresia. Eur J
Rare complex conditions require specialized knowledge and
Pediatr Surg 2015; 25(05):397–404
expertise in care and treatment,37 and patients with EA need 6 Nurminen P, Koivusalo A, Hukkinen M, Pakarinen M. Pneumonia
multidisciplinary follow-up care.38,39 It is known that 50 to after Repair of Esophageal Atresia-Incidence and Main Risk
75% of patients with EA consider their HRQOL restricted,23,40 Factors. Eur J Pediatr Surg 2018. Doi: 10.1055/s-0038-1675775

7 Rozensztrauch A, Smigiel R, Patkowski D. Congenital esophageal 24 Dellenmark-Blom M, Chaplin JE, Gatzinsky V, Jönsson L, Abra-
atresia-surgical treatment results in the context of quality of life. hamson K. Health-related quality of life among children, young
Eur J Pediatr Surg 2019; 29(03):266–270 people and adults with esophageal atresia: a review of the
8 Cella DF. Quality of life: concepts and definition. J Pain Symptom literature and recommendations for future research. Qual Life
Manage 1994;9(03):186–192 Res 2015;24(10):2433–2445
9 Flieder S, Dellenmark-Blom M, Witt S, et al. Generic health- 25 Jönsson L, Friberg LG, Gatzinsky V, Kötz K, Sillén U, Abrahamsson
related quality of life after repair of esophageal atresia and its K. Treatment and follow-up of patients with long-gap esophageal
determinants within a German-Swedish Cohort. Eur J Pediatr atresia: 15 years’ of experience from the Western Region of
Surg 2019; 29(01):75–84 Sweden. Eur J Pediatr Surg 2016;26(02):150–159
10 Legrand C, Michaud L, Salleron J, et al. Long-term outcome of 26 Khan KM, Krosch TC, Eickhoff JC, et al. Achievement of feeding
children with oesophageal atresia type III. Arch Dis Child 2012;97 milestones after primary repair of long-gap esophageal atresia.
(09):808–811 Early Hum Dev 2009;85(06):387–392
11 Lepeytre C, De Lagausie P, Merrot T, Baumstarck K, Oudyi M, 27 Baird R, Levesque D, Birnbaum R, Ramsay M. A pilot investigation
Dubus JC. [State of health, follow-up and medium-term outcome, of feeding problems in children with esophageal atresia. Dis
follow-up, and quality of life in children treated for type III Esophagus 2015;28(03):224–228
esophageal atresia]. Arch Pediatr 2013;20(10):1096–1104 28 Salö M, Stenström P, Anderberg M, Arnbjörnsson E. Anastomotic
12 Amin R, Knezevich M, Lingongo M, et al. Long-term quality of life strictures after esophageal atresia repair: timing of dilatation
in neonatal surgical disease. Ann Surg 2018;268(03):497–505 during the first two postoperative years. Surg J (N Y) 2018;4(02):
13 Peetsold MG, Heij HA, Deurloo JA, Gemke RJ. Health-related quality e62–e65
of life and its determinants in children and adolescents born with 29 Koivusalo AI, Pakarinen MP, Lindahl HG, Rintala RJ. Revisional
oesophageal atresia. Acta Paediatr 2010;99(03):411–417 surgery for recurrent tracheoesophageal fistula and anastomotic
14 Dingemann C, Meyer A, Kircher G, et al. Long-term health-related complications after repair of esophageal atresia in 258 infants.
quality of life after complex and/or complicated esophageal J Pediatr Surg 2015;50(02):250–254
atresia in adults and children registered in a German patient 30 Friedmacher F, Kroneis B, Huber-Zeyringer A, et al. Postoperative
support group. J Pediatr Surg 2014;49(04):631–638 complications and functional outcome after esophageal atresia
15 Dellenmark-Blom M, Chaplin JE, Gatzinsky V, et al. Health-related repair: results from Longitudinal Single-Center Follow-Up. J Gas-
quality of life experiences among children and adolescents born trointest Surg 2017;21(06):927–935
with esophageal atresia: development of a condition-specific ques- 31 Dellenmark-Blom M, Jonsson L, Gatzinsky V, et al. Clinical pre-
tionnaire for pediatric patients. J Pediatr Surg 2016;51(04):563–569 dictors and prevalence of receiving special preschool/school
16 Dellenmark-Blom M, Abrahamsson K, Quitmann JH, et al. Devel- support in children with repaired esophageal atresia. J Pediatr
opment and pilot-testing of a condition-specific instrument to Surg 2018;53(10):1970–1975
assess the quality-of-life in children and adolescents born with 32 Shah R, Varjavandi V, Krishnan U. Predictive factors for complica-
esophageal atresia. Dis Esophagus 2017;30(07):1–9 tions in children with esophageal atresia and tracheoesophageal
17 Dellenmark-Blom M, Dingemann J, Witt S, et al. The Esophageal- fistula. Dis Esophagus 2015;28(03):216–223
atresia-Quality-of-life questionnaires: feasibility, validity and 33 Castilloux J, Noble AJ, Faure C. Risk factors for short- and long-
reliability in Sweden and Germany. J Pediatr Gastroenterol Nutr term morbidity in children with esophageal atresia. J Pediatr
2018;67(04):469–477 2010;156(05):755–760
18 Witt S, Dellenmark-Blom M, Dingemann J, et al. Quality of life in 34 Drotar D. Measuring child health: scientific questions, challenges,
parents of children born with esophageal atresia. Eur J Pediatr and recommendations. Ambul Pediatr 2004;4(4, Suppl):353–357
Surg 2018. Doi: 10.1055/s-0038-1660867 35 Bullinger M. The challenge of cross-cultural quality of life assess-
19 Matza LS, Patrick DL, Riley AW, et al. Pediatric patient-reported ment. Psychol Health 1997;12(Suppl 6):815–825
outcome instruments for research to support medical product 36 Stoll C, Alembik Y, Dott B, Roth MP. Associated malformations in
labeling: report of the ISPOR PRO good research practices for the patients with esophageal atresia. Eur J Med Genet 2009;52(05):
assessment of children and adolescents task force. Value Health 287–290
2013;16(04):461–479 37 Wijnen R, Anzelewicz SM, Petersen C, Czauderna P. European
20 Benjamin K, Vernon MK, Patrick DL, Perfetto E, Nestler-Parr S, Burke Reference Networks: share, care, and cure-future or dream? Eur J
L. Patient-reported outcome and observer-reported outcome assess- Pediatr Surg 2017;27(05):388–394
ment in rare disease clinical trials: an ISPOR COA Emerging Good 38 IJsselstijn H, Gischler SJ, Toussaint L, Spoel M, Zijp MH, Tibboel D.
Practices Task Force Report. Value Health 2017;20(07):838–855 Growth and development after oesophageal atresia surgery: need
21 Wild D, Grove A, Martin M, et al; ISPOR Task Force for Translation for long-term multidisciplinary follow-up. Paediatr Respir Rev
and Cultural Adaptation. Principles of good practice for the transla- 2016;19:34–38
tion and cultural adaptation process for patient-reported outcomes 39 Krishnan U, Mousa H, Dall’Oglio L, et al. ESPGHAN-NASPGHAN
(PRO) measures: report of the ISPOR Task Force for Translation and Guidelines for the evaluation and treatment of gastrointestinal
Cultural Adaptation. Value Health 2005;8(02):94–104 and nutritional complications in children with esophageal atre-
22 Sullivan GM, Feinn R. Using effect size-or why the p value is not sia-tracheoesophageal fistula. J Pediatr Gastroenterol Nutr 2016;
enough. J Grad Med Educ 2012;4(03):279–282 63(05):550–570
23 Svoboda E, Fruithof J, Widenmann-Grolig A, et al. A patient led, 40 Hölscher AC, Laschat M, Choinitzki V, et al. Quality of life after
international study of long term outcomes of esophageal atresia: surgical treatment for esophageal atresia: long-term outcome of
EAT 1. J Pediatr Surg 2018;53(04):610–615 154 patients. Eur J Pediatr Surg 2017;27(05):443–448

Original Article
Novel Risk Score for Fetuses with Congenital

Diaphragmatic Hernia Based on Ultrasound Findings
Keita Terui1 Kouji Nagata2 Masahiro Hayakawa3 Hiroomi Okuyama4 Shoichirou Amari5
Akiko Yokoi6 Kouji Masumoto7 Naoto Urushihara8 Tadaharu Okazaki9 Noboru Inamura10
Katsuaki Toyoshima11 Keiichi Uchida12 Taizo Furukawa13 Manabu Okawada14 Yasunori Sato15
Noriaki Usui16
1 Department of Pediatric Surgery, Chiba University, Chiba, Japan Address for correspondence Keita Terui, MD, PhD, Department of
2 Department of Pediatric Surgery, Reproductive and Developmental Pediatric Surgery, Chiba University Graduate School of Medicine,
Medicine, Graduate School of Medical Sciences, Kyushu University, 1-8-1 Inohana, Chuo-ku, Chiba 260-8677, Japan (e-mail: [email protected]).
Fukuoka, Japan
3 Division of Neonatology, Center for Maternal-Neonatal Care,
Nagoya University Hospital, Nagoya, Japan
4 Department of Pediatric Surgery, Osaka University Graduate
School of Medicine, Suita, Osaka, Japan
5 Division of Neonatology, National Center for Child Health and
Development, Setagaya-ku, Tokyo, Japan
6 Department of Pediatric Surgery, Kobe Children’s Hospital, Kobe, Japan
7 Department of Pediatric Surgery, Tsukuba Daigaku, Tsukuba,
Ibaraki, Japan
8 Department of Pediatric Surgery, Shizuoka Children’s Hospital,
Shizuoka, Japan
9 Department of Pediatric Surgery, Juntendo University Urayasu
Hospital, Chiba, Japan
10 Department of Pediatrics, Kinki University, Higashiosaka, Osaka, Japan
11 Departments of Neonatology, Kanagawa Childrens Medical Center,
Yokohama, Kanagawa, Japan
12 Department of Gastrointestinal and Pediatric Surgery, Mie
University Graduate School of Medicine, Tsu, Mie, Japan
13 Department of Pediatric Surgery, Graduate School of Medical
Science, Kyoto Prefectural University of Medicine, Kyoto, Japan
14 Department of Pediatric General and Urogenital Surgery, Juntendo
University School of Medicine, Tokyo, Japan
15 Department of Preventive Medicine and Public Health, Keio
University, Minato-ku, Tokyo, Japan
16 Department of Pediatric Surgery, Osaka Women’s and Children’s
Hospital, Izumi, Japan
Eur J Pediatr Surg
Abstract Introduction We aimed to establish and validate a risk score for fetuses with
congenital diaphragmatic hernia (CDH) using only prenatal ultrasound findings.
Keywords Material and Methods Derivation (2011–2016, n ¼ 350) and validation (2006–2010,
► congenital n ¼ 270) cohorts were obtained from a Japanese CDH study group database. Using a
diaphragmatic hernia logistic regression analysis, we created a prediction model and weighted scoring
► fetus system from the derivation dataset and calculated the odds ratio of an unsatisfactory
► prognosis prognosis (death within 90 days of life or hospitalization duration exceeding 180 days).
► morbidity Five adverse prognostic factors obtained using prenatal ultrasound, including an

September 2, 2019
Fetuses with Congenital Diaphragmatic Hernia Terui et al.
observed/expected lung area-to-head circumference ratio (o/eLHR) <25%, liver herni-

ation occupying more than one-third of the thoracic space, thoracic stomach, right-side
CDH, and severe malformations, were used as predictors. The obtained model was
validated using the validation cohort.
Results The unsatisfactory prognosis prediction model was obtained based on the
adjusted odds ratios. The C statistics of the model were 0.83 and 0.80 in the derivation
and validation datasets, respectively. The five variables were weighted proportionally
to their adjusted odds ratios for an unsatisfactory prognosis (o/eLHR <25%, 1 point;
liver herniation occupying more than one-third of the thoracic space, 1 point; thoracic
stomach, 1 point; right-side CDH, 2 points; and severe malformations, 3 points).
Unsatisfactory prognosis rates for the low- (0–2 points), intermediate- (3–5 points),
and high-risk (6–8 points) groups were 17, 46, and 100%, respectively (p < 0.001), in
the validation cohort.
Conclusion Our simple risk score effectively predicted the prognosis of fetuses with
CDH.
Introduction Materials and Methods

Prenatally diagnosed congenital diaphragmatic hernia (CDH) Data and Patient Selection
remains one of the most challenging neonatal diseases in We used two databases from the Japanese CDH study group:
terms of appropriate management.1 Informing parents of the one from a nationwide survey that included 72 institutions
diagnosis of CDH during the prenatal period usually causes between 2006 and 201012 and one from a multicenter
great anxiety. Prenatal counseling is helpful for promoting registry that included 15 institutions between 2011 and
parents’ understanding of the condition and for planning 2016. The former was used as the validation dataset and
perinatal management.2 However, the spectrum of disease the latter was used as the derivation dataset in the risk
severity within CDH is broad.3 More accurate and straight- model. Definitions of all study variables are available in the
forward information regarding prognosis is imperative. Fur- original report.12 Patients with CDH who were diagnosed
thermore, parents of fetuses with CDH usually want to be prenatally and treated aggressively without palliative care
made aware of the morbidities as well as the potential for after birth were included in this study. Patients with chro-
mortality. Therefore, a simple system for predicting the mosomal abnormalities or bilateral CDH were excluded.
prognoses of morbidities is needed. Patients who underwent fetoscopic endoluminal tracheal
Numerous prenatal factors have been proposed as predic- occlusion (FETO) were also excluded.
tors of prognoses for CDH. Among these, fetal lung volume is a
major predictor. Measurements and standardization of fetal Ethical Approval
lung volume (observed/expected lung area-to-head circum- This study was conducted with the approval from the Ethics
ference ratio [o/eLHR]) via ultrasound are simple and widely Committee of Osaka University Hospital (approval number
used.4 The lung-to-thorax transverse area ratio (L/T ratio) is 11017) and the committees at all participating institutions.
another simple and reasonable method of evaluating fetal lung The requirement for signed informed consent was waived
volume,5 but its use is currently limited. Total lung volume because of the retrospective study design and the use of
measurements according to magnetic resonance imaging de-identified data. Details of the study were published on an
(MRI) have considerable predictive ability,6 but they lack institutional website, and individuals had the right to decline
versatility. Herniation of organs, such as the liver and stomach, participation in the study. The study was performed in
has been proposed as a powerful risk factor.7,8 Right-side accordance with the principles of the Declaration of Helsinki
herniation is thought to have a worse prognosis than and the Ethical Guidelines for Medical and Health Research
left-side herniation.9 Furthermore, severe malformations Involving Human Subjects.
influence the prognosis.10,11 However, few reports have com-
prehensively assessed these risk factors. For a more precise Outcome
prediction of the prognosis for CDH, a comprehensive assess- The primary outcome measure was unsatisfactory prognosis,
ment of prenatal risk factors is needed. which was defined by at least one of the following criteria:
Therefore, the aim of this study was to establish a simple death at 90 days of age or prolonged hospitalization exceed-
predictive risk score for fetuses with CDH, based solely on ing 180 days. The cutoff point, 180-day hospital stay, was set
prenatal ultrasound findings. based on a previous follow-up study.13

Variables
Results
The following risk factors identified on prenatal ultrasound
were used: o/eLHR <25%4; liver-up7; thoracic position of the Study Cohort
stomach8; right-side herniation9; and severe malforma- Of the 446 patients with CDH listed in the registry from 2011
tions.11 If the o/eLHR was measured more than once, then and 2016, 96 (22%) were excluded for the following reasons:
the minimum value was used. The patients were divided into no prenatal diagnosis (n ¼ 61; 14%), FETO (n ¼ 11; 2%),
two groups, high and low, according to the o/eLHR. The cutoff palliative care instead of aggressive treatment (n ¼ 15; 3%),
point of 25% was set based on the results of a previous and chromosomal abnormalities (n ¼ 19; 4%). No patients
report.14 Liver-up was defined as prenatally detected liver had bilateral CDH. Ultimately, the data of 350 patients were
herniation occupying more than one-third of the thoracic included in the derivation dataset. In the derivation dataset,
space.12 Severe malformations were limited to disorders that the following data were missing: o/eLHR for 39 (11%)
could be identified on prenatal ultrasound, such as cardiac patients; liver-up for 6 (2%) patients; and thoracic position
malformations, congenital central nervous system abnor- of the stomach for 12 (3%) patients. There were no missing
malities, and ventral anomalies. values associated with right-side herniation, severe malfor-
mations, or the outcome.
Statistical Model Building Of the 614 patients with CDH in the nationwide survey
To develop a model, a multivariable logistic regression from 2006 and 2010, 219 (36%) were excluded for the
analysis using the selected predictors was performed to following reasons: no prenatal diagnosis (n ¼ 172; 28%),
identify predictors of unsatisfactory prognosis and to esti- palliative care instead of aggressive treatment (n ¼ 21; 3%),
mate the adjusted odds ratios (OR) and 95% confidence bilateral hernia (n ¼ 4; 1%), chromosomal abnormalities
intervals (CI). (n ¼ 39; 6%), and lack of follow-up data at 90 days (n ¼ 8;
1%). No patients underwent FETO. Among the remaining 395
Risk Score Development patients, 125 (32%) were excluded because of the following
Each of the variables in the final model was assigned an missing predictive values: o/eLHR (113; 29%); liver-up (29;
integer-weighted score. The β coefficients were divided by 7%); and thoracic stomach (34; 9%). Ultimately, data from 270
the smallest absolute value of the regression coefficient and patients with available values for both predictors and out-
rounded to the nearest integer. Patients were separated into comes were included in the validation dataset.
three risk categories based on the score distribution. Baseline demographic, predictor, and outcome data of the
derivation and validation datasets are presented in ►Table 1.
Validation Apgar scores at 1 minute were higher (5 vs. 4; p ¼ 0.029) in
We calculated the area under the curve, that is, C-index, for the derivation dataset compared with those in the validation
the risk model to assess the model’s discriminatory power. dataset; however, patients in the derivation dataset had
The overall model performance was compared with that of lower o/eLHR (30.4 vs. 39.1; p < 0.001) and a higher propor-
each predictor alone using C-index for unsatisfactory prog- tion of large diaphragmatic defects (43 vs. 34%; p ¼ 0.041).
noses in the derivation dataset. The generalization ability of The rates of o/eLHR <25%, liver-up, thoracic stomach, right-
the best risk models was evaluated with the validation side CDH, and severe malformations in the derivation dataset
dataset. We calculated the C-index for all the best models were 32, 36, 70, 5, and 10%, respectively.
to predict unsatisfactory prognosis with this dataset. We also Ninety-six (27%) patients in the derivation and 65 (24%)
calculated the Hosmer–Lemeshow statistic of the risk model patients in the validation datasets (p ¼ 0.405) had unsatis-
using the derivation as well as the validation datasets. factory prognoses. Of the 96 patients with an unsatisfactory
To assess the proposed risk scoring system, morbidity and prognosis in the derivation dataset, 66 (69% of all patients
mortality were compared among the three groups in the with unsatisfactory prognoses) died within 90 days, and 30
validation dataset. The unsatisfactory prognosis rates were (31% of all patients with unsatisfactory prognoses and 10% of
compared between the derivation and validation datasets. the 90-day survivors) were hospitalized for more than
180 days. Among the 30 patients who required hospital stays
Statistical Analysis exceeding 180 days, 7 (23%) subsequently died; 5 of the 23
Statistical analyses were performed using JMP software survivors had cerebral palsy diagnosed after 1.5 years of life.
program version 13.01 (SAS Institute, Inc., Cary, North In comparison, among patients who were discharged within
Carolina, United States). Medians and interquartile ranges 180 days, only 1 of 254 (0.4%) died after 90 days of life and 2
were used to summarize ordinal scales and variables with of 222 (1%) had cerebral palsy. Only one patient who died
non-normal distributions. Means and standard deviations or between 91 and 180 days was assigned to not unsatisfactory
medians and interquartile ranges were used to summarize prognosis according to the definition.
continuous variables. In the univariate analyses, values were Severe malformations occurred in the derivation dataset
compared using the chi-square test. One-way analysis of as follows: severe cardiac anomalies (n ¼ 29; 83% of severe
variance with Tukey’s test was used to compare continuous malformations), including tetralogy of Fallot (n ¼ 7); hypo-
variables. The Kruskal–Wallis test was used to compare plastic left heart syndrome (n ¼ 5); double-outlet right ven-
ordinal scales and variables with non-normal distributions. tricle (n ¼ 4); single ventricle (n ¼ 4); neurological anomalies
p < 0.05 was regarded as statistically significant. (n ¼ 3; 9%); omphalocele (n ¼ 3; 9%).

Table 1 Demographic data of validation and derivation dataset
Validation dataset Derivation dataset p-Value

(n ¼ 270) (n ¼ 350)
Male sex, n (%) 147/270 (54) 185/350 (53) 0.726
Right-side hernia, n (%) 17/270 (6) 17/347 (5) 0.480
o/eLHR 39.1 (IQR 27.8–50.6) 30.4 (IQR 22.4–42.6) <0.001
o/eLHR <25%, n (%) 57/270 (21) 100/311 (32) 0.005
Liver-up, n (%) 79/270 (29) 124/344 (36) 0.181
Thoracic position of stomach, n (%) 203/270 (75) 235/338 (70) 0.122
Severe malformations, n (%) 14 / 270 (5) 35/350 (10) 0.172
Caesarian section, n (%) 201/270 (74) 255/349 (73) 0.782
Gestational age at birth (week) 37.7 (IQR 37.1–38.4) 37.7 (IQR 37.3–38.4) 0.564
Outborn, n (%) 0/270 (0) 3/350 (1) 0.252
Birth weight (g) 2,756 (IQR 2,532–2,960) 3,384 (IQR 2,990–3,940) 0.393
Apgar score at 1 min 4 (IQR 2–6) 5 (IQR 3–7) 0.029
Right to left shunt at DA, n (%) 127/258 (49) 140/328 (45) 0.273
Need for ECMO, n (%) 22/270 (8) 22/350 (6) 0.432
Need for prostaglandin I2, n (%) 53/266 (20) 53/350(15) 0.133
Need for iNO, n (%) 193/269 (72) 275/350 (79) 0.058
Surgery for CDH, n (%) 245/270 (91) 317/350 (91) 0.887
Defect size of the diaphragm, 83/244 (34) 126/294 (43) 0.041
C þ D, n (%)
Direct closure, n (%) 152/245 (62) 182/317 (57) 0.299
Unsatisfactory prognosis, n (%) 65/270 (24) 96/350 (27) 0.345
Death at 90 days of age, n (%) 50/270 (19) 66/350 (19) 0.915
Length of hospital stay >180 days, 15/220 (7) 30/284 (11) 0.144
n (%)a
Abbreviations: CDH, congenital diaphragmatic hernia; DA, ductus arteriosus; ECMO, extracorporeal membrane oxygenation; iNO, nitric oxide
inhalation; IQR, interquartile range; o/eLHR, observed-to-expected lung-to-head ratio.
a
Number of incidences was divided by number of 90-day survivors.
Derivation of the Model sum of the weighted predictors present in each case (o/eLHR
A multiple logistic regression analysis showed that thefollowing <25%, 1 point; liver-up, 1 point; thoracic stomach, 1 point;
five predictors: o/eLHR <25%, liver-up, thoracic stomach, right- right-side CDH, 2 points; and severe malformations, 3
side CDH, and severe malformations, were independent pre- points; ►Fig. 1). According to the observed unsatisfactory
dictors of an unsatisfactory prognosis: o/eLHR <25% (adjusted prognosis rates for each point (►Fig. 1), the patients were
OR, 3.4; 95% CI, 1.7–6.6; p < 0.001); liver-up (adjusted OR, divided into three groups: low-risk (0–2) group, intermedi-
3.0; 95% CI, 1.5–5.9; p ¼ 0.002); thoracic stomach (adjusted ate-risk (3–5) group, and high-risk (6–8) group.
OR, 2.8; 95% CI, 1.1–7.6; p ¼ 0.038); right-side CDH (adjusted OR,
4.9; 95% CI, 1.1–21.2; p ¼ 0.035); and severe malformations Validation of the Model and the Scoring System
(adjusted OR, 9.6; 95% CI, 3.7–24.7; p < 0.001) (►Table 2). The The c-statistics of the model in the derivation and validation
following equation was derived: datasets were 0.83 and 0.80, respectively, with Hosmer–
Probability of unsatisfactory prognosis ¼ 1–1/(1 þ eX) Lemeshow p-values of 0.917 and 0.461, respectively. In the
where X ¼ 0.43 þ 0.61 (o/eLHR <25%) þ 0.55 (liver- derivation dataset, our prediction model had better c-statis-
up) þ 0.52 (thoracic stomach) þ 0.79 (right CDH) þ 1.13 (severe tics than any of the single predictors (o/eLHR <25%, 0.71;
malformations). liver-up, 0.73; thoracic stomach, 0.61; right-side CDH, 0.52;
and severe malformations, 0.62; ►Fig. 2).
Risk Score Development The observed predictors, morbidity rates, and mortality
To make each predictor be weighted, the β coefficients were rates were compared among the low-risk, intermediate-risk,
divided by the smallest absolute value of the regression and high-risk groups in the validation dataset (►Table 3).
coefficient and rounded to the nearest integer. A risk scoring There were significant differences among the three groups
system for unsatisfactory prognoses was developed using the regarding surgery for CDH (p < 0.001), diaphragmatic defect

Table 2 Univariate and multivariate analyses for unsatisfactory prognosis
Variable N Univariate analysis Multivariate analysis Integer score assigned

(%)
Crude OR p-Value Adjusted OR p-Value
(95% CI) (95% CI)
o/eLHR <25% 100 (32) 6.3 <0.001 3.4 <0.001 1
(ref; 25%) (3.7–10.9) (1.7–6.6)
Liver-up 124 (36) 7.3 <0.001 3.0 0.002 1
(ref; No liver-up) (4.3–12.3) (1.5–5.9)
Thoracic position of stomach 235 (70) 3.2 <0.001 2.8 0.038 1
(ref; No thoracic position of stomach) (1.7–6.0) (1.1–7.6)
Right-side hernia, Yes 17 (5) 2.5 0.071 4.9 0.035 2
(ref: Left-side hernia) (0.9–6.6) (1.1–21.2)
Severe malformations, Yes 35 (10) 12.0 <0.001 9.6 <0.001 3
(ref: No severe malformation) (5.2–27.7) (3.7–24.7)
Abbreviations: CI, confidence interval; ref, reference; o/eLHR, observed-to-expected lung-to-head ratio; OR, odds ratio.
Fig. 1 Scoring system for the prediction of unsatisfactory prognoses based on prenatal ultrasound findings of fetuses with CDH (upper) and the
observed rates of unsatisfactory prognoses according to our scoring system in the derivation dataset (lower). CDH, congenital diaphragmatic
hernia; o/eLHR, observed/expected lung area-to-head circumference ratio.
C þ D (p < 0.001), direct closure (p < 0.001), unsatisfactory Discussion

prognosis (p < 0.001), and death at 90 days of age (p < 0.001).
The observed rates of unsatisfactory prognoses in the low- The principal findings of this study in which a new scoring
risk, intermediate-risk, and high-risk groups were 12, 56, and system specific to fetuses with CDH was created and
100%, respectively, in the derivation dataset (p < 0.001), and validated were as follows: an unsatisfactory prognosis
17, 46, and 100%, respectively, in the validation dataset can be clearly predicted using five clinical variables that
(p < 0.001). are routinely visible on prenatal ultrasound; severe

Fig. 2 Comparison of c-statistics between our model and single predictors in the derivation dataset. CDH, congenital diaphragmatic hernia;
o/eLHR, observed/expected lung area-to-head circumference ratio.
Table 3 Comparison of the three groups of the risk stratification system in validation dataset
Low-risk group Intermediate-risk group High-risk group p-Value

(n ¼ 216) (n ¼ 48) (n ¼ 6)
Male sex, n (%) 117 (54) 27 (56) 3 (50) 0.943
Right-side hernia, n (%) 2 (1) 14 (29) 1 (17) <0.001
o/eLHRa,b 41 (IQR 32–51) 24 (IQR 19–43) 15 (IQR 13–29) <0.001
o/eLHR <25%, n (%) 25 (12) 27 (56) 5 (83) <0.001
Liver-up, n (%) 32 (15) 41 (85) 6 (100) <0.001
Thoracic position of stomach, n (%) 162 (75) 36 (75) 5 (83) 0.888
Severe malformation, n (%) 0 (0) 8 (17) 6 (100) <0.001
Gestational age at birth (week) 37.7 37.6 37.5 ns
(IQR 37.1–38.3) (IQR 37.0–38.7) (IQR 36.4–38.1)
Birth weight (g) 2,772 2,660 2399 ns
(IQR 2,554–2,968) (IQR 2,442–2,982) (IQR 2,069–2,626)
Apgar score at 1 minute§ 4 (IQR 3–6) 3 (IQR 2–5) 3 (IQR 2–5) 0.013
Need for ECMO, n (%) 14 (7) 7 (15) 1 (17) 0.173
Surgery for CDH, n (%) 207 (96) 36 (75) 2 (33) <0.001
Diaphragmatic defect C þ D, n (%) 55 (27) 26 (72) 2 (100) <0.001
Direct closure, n (%) 143 (69) 9 (26) 0 (0) <0.001
Unsatisfactory prognosis, n (%) 37 (17) 22 (46) 6 (100) <0.001
Death at 90 days of age, n (%) 27 (13) 18 (38) 5 (83) <0.001
c
Length of hospital stay >180 days, n (%) 11 (5) 4 (8) 1 (17) 0.453
Abbreviations: CDH, congenital diaphragmatic hernia; ECMO, extracorporeal membrane oxygenation; iNO, nitric oxide inhalation; IQR, interquartile
range; ns, not significant; o/eLHR, observed-to-expected lung-to-head ratio.
a
p < 0.001 Low- vs. intermediate-risk group.
b
p < 0.05 low- vs. high-risk group and intermediate- vs. high-risk group.
c
Number of incidences was divided by number of 90-day survivors.
malformations were the most powerful independent pre- During prenatal counseling for CDH, several known pre-
dictors of an unsatisfactory prognosis; and our model dictors are often used to share information regarding the fetal
provided better discrimination of unsatisfactory prognoses prognosis, including o/eLHR, positions of the liver and stom-
than did the single predictors. ach, herniation location (left or right side), and associated

malformations.2 However, it was unclear which factor was liver-up and the thoracic position of the stomach among
most powerful or which factors depended on other factors. patients with right-side CDH. Our model, which can be used
Furthermore, the positions of the liver and stomach are to evaluate patients with either left-side or right-side CDH, was
specific risk factors for left-side herniation only and cannot well-adjusted for better prediction.
be used for right-side herniation. Therefore, based solely on The selection of the primary outcome in this study was
their experiences, counselors have chosen to use variable unconventional. Unsatisfactory prognosis, which included
factors among several predictors. The present scoring system death at 90 days of age and prolonged hospitalization >180
provides a comprehensive method of evaluating the prenatal days, was used instead of more commonly used outcomes such
information of fetuses with CDH. These findings are useful to as survival rate.1,4,5,7–9,15–20 This is because the survival rate is
enhance the parents’ understanding during prenatal counsel- not the only information that parents of fetuses with CDH need
ing and for planning perinatal management. to know. Another focus of attention for parents is the presence
Several predictive equations have been proposed based on or absence of severe morbidity.2 Patients with CDH may have a
multiple risk factors. The CDH study group created a logistic wide range of morbidities, such as severe pulmonary hyper-
equation that used birth weight and the 5-minute Apgar tension, respiratory distress, nutritional abnormalities, and
score.15 The Wilford Hall/Santa Rosa group provided a pre- neurological disorders.3 To integrate various morbidities into a
diction formula that was based on the gap between the single outcome, the duration of hospitalization was added as
highest PaO2 and the highest PCO2 of arterial blood gas values an alternative to actual morbidity.3 This was based on the
obtained during the initial 24 hours of life.16 The Canadian fact that the higher rates of morbidities were correlated
Neonatal Network developed a method based on the combi- with longer durations of hospitalization.22 Furthermore, the
nation of the Score for Neonatal Acute Physiology version II length of hospital stay was found to be the only factor
(SNAP-II) and gestational age.17 Additionally, stratification associated with a lower quality of life.23
systems based on two major risk factors were provided by The current study had several limitations. First, the ther-
the CDH study group (size of diaphragmatic defect and apeutic strategies were not uniform because of the multi-
major cardiac anomaly)18 and the Japanese CDH study group center study design. Second, the severe malformations factor
(1-minute Apgar score and oxygenation index).19 All these was strongly affected by the judgment of the attending
predictive systems were based on postnatal information. To physician, who determined whether the fetal malformations
the best of our knowledge, the present scoring system is the were severe. The significance of techniques to detect specific
first to use only information obtained prenatally. prenatal findings and make accurate diagnoses is increasing.
One of the purposes of the present study was to develop a Third, this scoring system cannot be applied to patients with
model that can distinguish the most severe cases, which are bilateral CDH or patients with chromosomal abnormalities
difficult to cure. If the probability of death is nearly 100%, the because they were eliminated from this cohort. Aneuploidy
prognostic prediction must be informative for both parents identified after birth invalidates this prenatal prediction
and prenatal counselors. In our scoring system, the rate of method. Fourth, the patients in the derivation dataset had
unsatisfactory prognosis of high-risk group was 100% in the more severe disease than those in the validation dataset
validation dataset, even though this group had very small because the derivation dataset was originally based on the
number of cases; only 6 out of 270 patients (2.2%) in the registrations from mainly high-volume centers; the valida-
validation dataset. Thus, when patients in the high-risk tion dataset was derived from a nationwide study that
group are encountered, appropriate counseling that presents included both high-volume and low-volume centers. Finally,
palliative care as an option should be provided. although the present model was validated in another cohort,
This risk model is necessary to accurately predict the that validation was also performed in a cohort of Japanese
prognosis of CDH, and it should be used in clinical practice. patients. External validation in an independent international
Therefore, readily available and widely used prenatal ultra- sample is required to confirm the prognostic value and broad
sound findings provided the optimal parameters for developing applicability of the scoring system.
the risk score equation. Among the ultrasound findings, the In conclusion, we have presented a simple risk scoring
pulmonary artery diameter and vascularization index system based on widely available prenatal ultrasound find-
have great predictive ability,20 but they were not used because ings. Our scoring system is capable of predicting unsatisfac-
the measurements are not easy to perform and are not tory prognoses for fetuses with CDH.
sufficiently standardized. Assessing prenatal MRI findings is
also a powerful method of evaluating total lung volume,21 but Funding
these findings were not used as variables because the evalua- This study was funded by the Ministry of Health, Labor,
tion method is neither simple nor well-standardized. and Welfare of Japan (Grant/Award Number: “H24-Nan-
Among the five selected variables, severe malformations chi-Ippan-034”).
and right-side hernia were the most and second-most power-
ful independent predictors of unsatisfactory prognosis, respec- Conflict of interest
tively. Interestingly, right-side hernia was not a significant risk T.F., M.H., N.I., K.M., K.N., M.O., T.O., H.O., Y.S., K.T., K.T., K.
factor in the univariate analysis, but it was significant in the U., N.U., N.U., A.Y., and S.A. report grants from the Ministry
multivariate analysis. This mismatch probably resulted from of Health, Labor and Welfare of Japan, during the conduct
statistical adjustments based on the low predictive ability of of the study.

Acknowledgment 12 Nagata K, Usui N, Kanamori Y, et al. The current profile and

The authors gratefully acknowledge all institutions that outcome of congenital diaphragmatic hernia: a nationwide sur-
collected the data used for the present study. vey in Japan. J Pediatr Surg 2013;48(04):738–744
13 Terui K, Nagata K, Hayakawa M, et al. Growth assessment and the
risk of growth retardation in congenital diaphragmatic hernia: a
long-term follow-up study from the Japanese Congenital Dia-
References phragmatic Hernia Study Group. Eur J Pediatr Surg 2016;26(01):
1 Skari H, Bjornland K, Haugen G, Egeland T, Emblem R. Congenital 60–66
diaphragmatic hernia: a meta-analysis of mortality factors. 14 Snoek KG, Reiss IK, Greenough A, et al; CDH EURO Consortium.
J Pediatr Surg 2000;35(08):1187–1197 Standardized postnatal management of infants with congenital
2 Berman L, Jackson J, Miller K, Kowalski R, Kolm P, Luks FI. Expert diaphragmatic hernia in Europe: the CDH EURO Consortium
surgical consensus for prenatal counseling using the Delphi Consensus—2015 Update. Neonatology 2016;110(01):66–74
method. J Pediatr Surg 2018;53(08):1592–1599 15 Congenital Diaphragmatic Hernia Study Group. Estimating dis-
3 Puligandla PS, Grabowski J, Austin M, et al. Management of ease severity of congenital diaphragmatic hernia in the first
congenital diaphragmatic hernia: a systematic review from the 5 minutes of life. J Pediatr Surg 2001;36(01):141–145
APSA outcomes and evidence based practice committee. J Pediatr 16 Schultz CM, DiGeronimo RJ, Yoder BA; Congenital Diaphragmatic
Surg 2015;50(11):1958–1970 Hernia Study Group. Congenital diaphragmatic hernia: a simpli-
4 Senat MV, Bouchghoul H, Stirnemann J, et al; Center for Rare fied postnatal predictor of outcome. J Pediatr Surg 2007;42(03):
Diseases: Congenital Diaphragmatic Hernia. Prognosis of isolated 510–516
congenital diaphragmatic hernia using lung-area-to-head-cir- 17 Richardson DK, Corcoran JD, Escobar GJ, Lee SK. SNAP-II and
cumference ratio: variability across centers in a national perinatal SNAPPE-II: simplified newborn illness severity and mortality
network. Ultrasound Obstet Gynecol 2018;51(02):208–213 risk scores. J Pediatr 2001;138(01):92–100
5 Usui N, Kitano Y, Okuyama H, et al. Reliability of the lung to thorax 18 Lally KP, Lasky RE, Lally PA, et al; Congenital Diaphragmatic
transverse area ratio as a predictive parameter in fetuses with Hernia Study Group. Standardized reporting for congenital dia-
congenital diaphragmatic hernia. Pediatr Surg Int 2011;27(01):39–45 phragmatic hernia—an international consensus. J Pediatr Surg
6 Ruano R, Lazar DA, Cass DL, et al. Fetal lung volume and quantifi- 2013;48(12):2408–2415
cation of liver herniation by magnetic resonance imaging in 19 Terui K, Nagata K, Kanamori Y, et al. Risk stratification for
isolated congenital diaphragmatic hernia. Ultrasound Obstet congenital diaphragmatic hernia by factors within 24 h after
Gynecol 2014;43(06):662–669 birth. J Perinatol 2017;37(07):805–808
7 Kitano Y, Nakagawa S, Kuroda T, et al. Liver position in fetal congenital 20 Ruano R, Takashi E, da Silva MM, Campos JA, Tannuri U, Zugaib M.
diaphragmatic hernia retains a prognostic value in the era of lung- Prediction and probability of neonatal outcome in isolated congen-
protective strategy. J Pediatr Surg 2005;40(12):1827–1832 ital diaphragmatic hernia using multiple ultrasound parameters.
8 Kitano Y, Okuyama H, Saito M, et al. Re-evaluation of stomach Ultrasound Obstet Gynecol 2012;39(01):42–49
position as a simple prognostic factor in fetal left congenital 21 Neff KW, Kilian AK, Schaible T, Schütz EM, Büsing KA. Prediction
diaphragmatic hernia: a multicenter survey in Japan. Ultrasound of mortality and need for neonatal extracorporeal membrane
Obstet Gynecol 2011;37(03):277–282 oxygenation in fetuses with congenital diaphragmatic hernia:
9 Duess JW, Zani-Ruttenstock EM, Garriboli M, Puri P, Pierro A, logistic regression analysis based on MRI fetal lung volume
Hoellwarth ME. Outcome of right-sided diaphragmatic hernia measurements. AJR Am J Roentgenol 2007;189(06):1307–1311
repair: a multicentre study. Pediatr Surg Int 2015;31(05):465–471 22 Putnam LR, Harting MT, Tsao K, et al; Congenital Diaphragmatic
10 Brindle ME, Cook EF, Tibboel D, Lally PA, Lally KP; Congenital Hernia Study Group. Congenital diaphragmatic hernia defect size
Diaphragmatic Hernia Study Group. A clinical prediction rule for and infant morbidity at discharge. Pediatrics 2016;138(05):
the severity of congenital diaphragmatic hernias in newborns. e20162043
Pediatrics 2014;134(02):e413–e419 23 Peetsold MG, Huisman J, Hofman VE, Heij HA, Raat H, Gemke RJ.
11 Gentili A, Pasini L, Iannella E, et al. Predictive outcome indexes in Psychological outcome and quality of life in children born with
neonatal congenital diaphragmatic hernia. J Matern Fetal Neona- congenital diaphragmatic hernia. Arch Dis Child 2009;94(11):
tal Med 2015;28(13):1602–1607 834–840

@medicinejournal European Journal of Pediatric Surgery January 2020

Uploaded by

Copyright:

Available Formats

@medicinejournal European Journal of Pediatric Surgery January 2020

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

@medicinejournal European Journal of Pediatric Surgery January 2020

Uploaded by

Copyright:

Available Formats

Original Article

Potential Beneﬁts of Laparoscopic Repair of

Eur J Pediatr Surg

received © Georg Thieme Verlag KG DOI https://doi.org/

Perioperative Factors and Outcome Perioperative Data and Outcome

European Journal of Pediatric Surgery

Table 1 Preoperative patient characteristics Table 2 Perioperative data and outcome

Primary open Laparoscopic p-Value Primary open Laparoscopic p-Value

European Journal of Pediatric Surgery

European Journal of Pediatric Surgery

European Journal of Pediatric Surgery

European Journal of Pediatric Surgery

“Tuft Cells”: A New Player in Hirschsprung’s

Eur J Pediatr Surg

Introduction been shown to play a key role in gastrointestinal chemo-

received © Georg Thieme Verlag KG DOI https://doi.org/

European Journal of Pediatric Surgery

Quantitative Real-Time Polymerase Chain Reaction

European Journal of Pediatric Surgery

Discussion tracts of mice express an incomplete neuronal-type cholin-

European Journal of Pediatric Surgery

European Journal of Pediatric Surgery

Management of Spontaneous Pneumothorax in

Abstract Introduction Management of primary spontaneous pneumothorax (PSP) is mainly

Introduction The incidence of PSP in the pediatric population is low and

received © Georg Thieme Verlag KG DOI https://doi.org/

European Journal of Pediatric Surgery

Quality Assessment 100% were considered as low, moderate, substantial, and of

Fig. 1 Diagram of workﬂow in the systematic review and meta-analysis.

European Journal of Pediatric Surgery

European Journal of Pediatric Surgery

European Journal of Pediatric Surgery

Size of Pneumothorax The higher incidence of PSP among neonates (mainly

European Journal of Pediatric Surgery

Table 3 Methods of quantiﬁcation of pneumothorax

Author Year Pneumothorax quantification

European Journal of Pediatric Surgery

European Journal of Pediatric Surgery

11. Baseline equivalence 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2

Quality assessment No. of patients Effect Quality

Very low quality: We are very uncertain about the estimate.

European Journal of Pediatric Surgery

European Journal of Pediatric Surgery

European Journal of Pediatric Surgery

Outcomes of Primary versus Multiple-Staged

received © Georg Thieme Verlag KG DOI https://doi.org/

European Journal of Pediatric Surgery

All patients admitted All patients operated (N ¼ 1,137) p-Value

Abbreviation: NS, not signiﬁcant.

European Journal of Pediatric Surgery

European Journal of Pediatric Surgery

Readmission More cLOS

Abbreviations: CI, conﬁdence interval; cLOS, cumulative length of hospital stay.

European Journal of Pediatric Surgery

European Journal of Pediatric Surgery

European Journal of Pediatric Surgery

Indications, Surgical Complications, and Long-Term

Eur J Pediatr Surg

received © Georg Thieme Verlag KG DOI https://doi.org/